A  B  C  D  E  F  G  H  I  J  K  L  M  N  O  P  Q  R  S  T  U  V  W  X  Y  Z  

  Subjects -> ELECTRONICS (Total: 207 journals)
The end of the list has been reached or no journals were found for your choice.
   Most Followed Journals

   Top Publishers

   Top Subjects
Similar Journals
Journal Cover
IEEE Transactions on Geoscience and Remote Sensing
Journal Prestige (SJR): 2.649
Citation Impact (citeScore): 6
Number of Followers: 213  
 
  Hybrid Journal Hybrid journal (It can contain Open Access articles)
ISSN (Print) 0196-2892 - ISSN (Online) 1558-0644
Published by IEEE Homepage  [228 journals]
  • Bistatic Radar Tomography of Shear Margins: Simulated Temperature and
           Basal Material Inversions

    • Free pre-print version: Loading...

      Authors: Nicole Bienert;Dustin M. Schroeder;Paul Summers;
      Pages: 1 - 16
      Abstract: A significant portion of the sea level contributions of Antarctica and Greenland comes from ice streams, but the physical processes controlling ice stream width are poorly understood, especially when topographic controls are absent. Recent modeling studies have indicated that ice stream width may be controlled by elevated temperatures inside ice stream shear margins. While radio echo sounders can provide measurements of englacial water storage and subglacial conditions, existing radar-sounding techniques cannot measure temperature profiles at the scale required to test this hypothesis. We propose using a wide-angle radar survey and tomographic inversion to resolve temperature profiles, gradients, and anomalies at the scale required to study the thermophysical controls on shear margins. Recent work produced a bistatic radar system capable of obtaining the long offsets required for well-constrained inversions; however, shear-margin-specific temperature inversion techniques have not been developed for this system. In this article, we develop Newton’s method and alternating direction method of multipliers’ inversions for estimating temperature distribution and basal material across ice stream shear margins. We evaluate the performance of these inversion techniques on simulated bistatic radar-sounding data. Our results suggest that bistatic radar tomography experiments should be able to produce temperature maps on 50 m $times50$ m grids with $0.83~^{circ} text{C}~pm ~0.084~^{circ} text{C}$ mean temperature error, $3.58~^{circ} text{C}~pm ~0.20~^{circ} text{C}$ maximum temperature error, and an error in relative basal permittivity of $0.63~pm ~0.08$ for a 4-km transect.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • IR Digital Holography for Remote Sensing of Structures

    • Free pre-print version: Loading...

      Authors: Eugenio Pugliese;Massimiliano Locatelli;Flavio Bocchi;Pasquale Poggi;David Jafrancesco;Giuseppe Falzone;Nicola Signorini;Daniele Spina;Riccardo Meucci;
      Pages: 1 - 7
      Abstract: In this article, we present a comparison between the data obtained by infrared digital holography (IRDH) and conventional accelerometers in order to evaluate the structural stability of buildings. The tests have carried out on historical buildings located in Tuscany (Italy) under the continuous surveillance of the Italian Civil Protection Department (DPC) by means of accelerometers networks. From the comparison, it emerges that the holographic technique can be profitably used in the field of dynamical analysis of structures thanks to its high sensitivity and to the possibility to operate from remote. Indeed, the holographic technique is able to detect micromovements of the order of $mathbf {1/100}$ of the used laser wavelength at $mathbf {10.6 ~mu {mathrm{ m}}}$ without entering the structures.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • 2-D Modeling and Analysis of Time-Domain Electromagnetic Anomalous
           Diffusion With Space-Fractional Derivative

    • Free pre-print version: Loading...

      Authors: Yibing Yu;Quanming Gao;Xuejiao Zhao;Yanju Ji;
      Pages: 1 - 13
      Abstract: Recently, the electromagnetic (EM) anomalous diffusion phenomenon has been observed in time-domain EM (TDEM) surveys. Furthermore, the data interpretation accuracy has been reduced by adopting the traditional EM theory and methods. A number of models, such as random medium and roughness electrical conductivity theory, have been adopted to model the EM anomalous diffusion. However, problems such as modeling difficulty and massive discretization exist regarding characterizing the long-range correlation of EM anomalous diffusion. The space-fractional derivative has been proven to preferably describe the long-range correlation characteristic. Only a handful of studies on TDEM anomalous diffusion with space-fractional derivative have been conducted due to the difficulties in computational engineering problems. Therefore, we performed a series of studies about 2-D TDEM anomalous diffusion with space-fractional derivative. The 2-D TDEM space-fractional diffusion equation was constructed based on the space-fractional Ohm’s law model. Furthermore, the discretization and iteration forms of the control equation were derived based on the finite element method (FEM) by introducing the Riemann–Liouville (R–L)-type Riesz fractional derivatives. The 2-D mountain-shaped function and partial integration method (PIM) were combined to convert the fractional derivative into the primitive function form. Hence, the 2-D modeling of the space-fractional EM diffusion was realized. The effectiveness of our method was verified by the function construction method and wavenumber-domain analytical solution. The spatial and temporal characteristics of the space-fractional EM diffusion were analyzed by different geological models. Furthermore, we discuss the differences with the classical EM diffusion. Our method can effectively model the space-fractional EM diffusion in TDEM surveys and provide theoretical bases for improving the TDEM inter-retation accuracy with complex geological conditions.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Method for Separating the O/X Mode Signals in Vertical Ionograms Based
           on Improved U-Shaped Encoder–Decoder Network

    • Free pre-print version: Loading...

      Authors: Hongchun Li;Chengfeng Zhang;Bo Yin;Xiaoyi Jia;Jiali Xu;Jiawei Xue;Mengfei Ma;
      Pages: 1 - 13
      Abstract: The accuracy of O/X mode separation in vertical ionograms directly determines the quality of pattern discrimination results and metrics, which is of great significance to ionospheric research. It is extremely complicated to separate the O/X mode of the vertical ionograms because of the environmental noise, interference, and the time-varying dispersion characteristics of the ionosphere itself. In this article, we propose a method for separating the O/X mode signal in vertical ionograms based on an improved U-shaped encoder–decoder network, named vertical ionogram separation U-shaped network (VIS-UNet). Our model is based on the encoder–decoder architecture. It introduces the residual convolution to avoid network performance degradation and utilizes the attention module to improve the attention of the signal characteristic. In addition, we design an adaptive loss function to expedite the convergent speed of the model. Experimental results show that our model performs better than the baselines for the task of the O/X mode signal separation: 1) the method in this article has low requirements on the vertical ionospheric sounding system and the ionograms obtained by the single-channel vertical ionospheric sounding system can realize the separation of O/X mode signal at the pixel level; 2) it has strong universality and is insensitive to the signal integrity and the ionospheric pattern of the vertical ionograms; and 3) it performs better for the separation task. The mean intersection over union (MIOU) of the O/X mode separation task reaches 91.97% and the performance is significantly improved compared with the existing methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Asymmetric Cross-Attention Hierarchical Network Based on CNN and
           Transformer for Bitemporal Remote Sensing Images Change Detection

    • Free pre-print version: Loading...

      Authors: Xiaofeng Zhang;Shuli Cheng;Liejun Wang;Haojin Li;
      Pages: 1 - 15
      Abstract: As an important task in the field of remote sensing (RS) image processing, RS image change detection (CD) has made significant advances through the use of convolutional neural networks (CNNs). The transformer has recently been introduced into the field of CD due to its excellent global perception capabilities. Some works have attempted to combine CNN and transformer to jointly harvest local-global features; however, these works have not paid much attention to the interaction between the features extracted by both. Also, the use of the transformer has resulted in significant resource consumption. In this article, we propose the Asymmetric Cross-attention Hierarchical Network (ACAHNet) by combining CNN and transformer in a series-parallel manner. The proposed Asymmetric Multiheaded Cross Attention (AMCA) module reduces the quadratic computational complexity of the transformer to linear, and the module enhances the interaction between features extracted from the CNN and the transformer. Different from the early and late fusion strategies employed in previous work, the effectiveness of the mid-term fusion strategy employed by ACAHNet shows a new choice of timing for feature fusion in the CD task. Our experiments on the proposed method on three public datasets show that our network has a better performance in terms of effectiveness and computational resource consumption compared to other comparative methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multiple-Space Deep Learning Schemes for Inverse Scattering Problems

    • Free pre-print version: Loading...

      Authors: Yusong Wang;Zheng Zong;Siyuan He;Zhun Wei;
      Pages: 1 - 11
      Abstract: Recently, deep learning methods have made significant success in inverse scattering problems (ISPs). However, learning approaches that work in different spaces, such as frequency and real space, are seldom explored in solving ISPs. In this work, multiple-space deep learning schemes (MSDLSs) incorporating frequency-space and real-space processing are studied. Specifically, a network that works in low-frequency subspace is first introduced. Then, serial MSDLSs are introduced by combining frequency-space and real-space networks in a serial way to enable networks in different spaces work complementarily. Finally, to further enable dynamic interaction between multiple-space information during both training and testing stages, a parallel MSDLS is proposed. The proposed MSDLSs are presented under the framework of backpropagation scheme (BPS). It is shown by synthetic and experimental tests that the MSDLSs have a consistent improvement over BPS. It is expected that the proposed schemes will find applications on other inverse problems where an incorporation of multiple-space information is needed.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Three-Dimensional Ray-Based Tomographic Approach for Contactless GPR
           Imaging

    • Free pre-print version: Loading...

      Authors: Gianluca Gennarelli;Carlo Noviello;Giovanni Ludeno;Giuseppe Esposito;Francesco Soldovieri;Ilaria Catapano;
      Pages: 1 - 14
      Abstract: This article proposes a 3-D imaging approach for contactless ground penetrating radar surveys. The imaging problem is formulated in the linear inverse scattering context and solved by using the singular values decomposition tool. A ray-based model accounting for the electromagnetic (EM) signal propagation into an inhomogeneous medium is developed to accurately evaluate the kernel of the integral equation to be inverted. Under the proposed model, an analysis of the spatial resolution performance is carried out as a function of the geometrical and EM parameters of the scenario. To this end, theoretical concepts based on diffraction tomography and the singular value decomposition of the scattering operator are exploited. Reconstruction results based on full-wave simulated data assess the feasibility of the imaging approach.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Comprehensive Emission Model for Layered Irregular and Inhomogeneous
           Medium

    • Free pre-print version: Loading...

      Authors: Dongjin Bai;Xiaolong Dong;Saibun Tjuatja;Di Zhu;Zijin Zhang;
      Pages: 1 - 20
      Abstract: Reported radiometric measurements indicate that incoherent and coherent radiative transfer processes characterize the microwave emission features of layered irregular and inhomogeneous medium. In order to develop a forward emission model to be capable of considering the medium and boundary scattering and coherent boundary interaction that may exist in general layered medium, in this article, we present a comprehensive layer emission model (CLEM) based on the scattering operator formulation. We introduce wave-based coherent multiple reflection operators to account for coherent boundary interaction and first integrate them with the intensity-based multiple scattering processes, allowing the comprehensive description of rough boundary scattering, volume scattering, medium/boundary interaction, and coherent boundary interaction in the framework of CLEM. Simulations and analyses for ice- and snow-covered ground cases are conducted based on CLEM to evaluate the coherent boundary interaction and different impacting factors. Validation on CLEM is conducted with emission observations of snow-covered terrain in campaign Nordic Snow Radar Experiment (NoSREx) 2010–2013 during dry snow period and time-series emission observations of frozen soil during freezing and thawing processes. CLEM simulation results show a good agreement with measurements. For NoSREx, root-mean-square errors (RMSEs) at L- to Ka-band are below 5.5 K for both polarizations, which are of different levels of promotion compared with the incoherent model simulations, especially for horizontal polarization at L- and X-band. Application to frozen soil case illustrates the capability of CLEM to explain coherent oscillation feature with impact from incoherent scattering effect.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Spatiotemporal Interpolation Graph Convolutional Network for Estimating
           PM₂.₅ Concentrations Based on Urban Functional Zones

    • Free pre-print version: Loading...

      Authors: Xinya Chen;Yinghua Zhang;Yuebin Wang;Liqiang Zhang;Zhiyu Yi;Hanchao Zhang;P. Takis Mathiopoulos;
      Pages: 1 - 14
      Abstract: Urban functional zones (UFZs) contain abundant landscape information that can be adopted to better understand the surroundings. Various landscape compositions and configurations reflect different human activities, which may affect the particulate matter (PM2.5) concentrations. The very high-resolution (VHR) image features can reflect the physical and spatial structures of the UFZs. However, the existing PM2.5 estimation methods neither have been based on the scale of UFZs, nor have the VHR image features of UFZs as independent variables. Hence, this article proposes a spatiotemporal interpolation graph convolutional network (STI-GCN) model and introduces VHR image features to achieve PM2.5 estimation in UFZs. First, UFZs are split, and VHR image features are extracted by the visual geometry group 16 (VGG16). Subsequently, meteorological factors, aerosol optical depth (AOD), and VHR image features are used to estimate the PM2.5 concentrations at the scale of the UFZs. The two metropolises, Beijing and Shanghai, are chosen to assess the validity of the STI-GCN model. As for Beijing and Shanghai, the overall accuracy ${R^{2}}$ of the STI-GCN model can reach 0.96 and 0.89, the root-mean-square errors (RMSEs) are 8.15 and 6.40 $mu text {g}/{text {m}^{3}}$ , the mean absolute errors (MAEs) are 5.51 and 4.78 $mu text {g}/{text {m}^{3}}$ , and the relative prediction errors (RPEs) are 18.53% and 17.38%, respectively. Experiments show that the STI-GCN consistently outperforms other models. What’s more, the PM2.5 values are relatively high in commercial and official zones (COZs) and relatively low in urban green zones (UGZs).
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Absorbing Aerosol Optical Depth From OMI/TROPOMI Based on the GBRT
           Algorithm and AERONET Data in Asia

    • Free pre-print version: Loading...

      Authors: Ding Li;Jason Blake Cohen;Kai Qin;Yong Xue;Lanlan Rao;
      Pages: 1 - 10
      Abstract: Quantifying the concentration of absorbing aerosol is essential for pollution tracking and calculation of atmospheric radiative forcing. To quickly obtain absorbing aerosol optical depth (AAOD) with high-resolution and high-accuracy, the gradient boosted regression trees (GBRT) method based on the joint data from Ozone Monitoring Instrument (OMI), Moderate Resolution Imaging Spectro-Radiometer (MODIS), and AErosol RObotic NETwork (AERONET) is used for TROPOspheric Monitoring Instrument (TROPOMI). Compared with the ground-based data, the correlation coefficient of the results is greater than 0.6 and the difference is generally within ±0.04. Compared with OMI data, the underestimation has been greatly improved. By further restricting the impact factors, three valid conclusions can be drawn: 1) the model with more spatial difference information achieves better results than the model with more temporal difference information; 2) the training dataset with a high cloud fraction (0.1–0.4) can partly improve the performance of GBRT results; and 3) when aerosol optical depth (AOD) is less than 0.3, the perform of retrieved AAODs is still good by comparing with ground-based measurements. The novel finding is expected to contribute to regional and even urban anthropogenic pollution research.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Relative Importance of Radar Variables for Nowcasting Heavy Rainfall: A
           Machine Learning Approach

    • Free pre-print version: Loading...

      Authors: Yi Victor Wang;Seung Hee Kim;Geunsu Lyu;Choeng-Lyong Lee;Gyuwon Lee;Ki-Hong Min;Menas C. Kafatos;
      Pages: 1 - 14
      Abstract: Highly short-term forecasting, or nowcasting, of heavy rainfall due to rapidly evolving mesoscale convective systems (MCSs) is particularly challenging for traditional numerical weather prediction (NWP) models. To overcome such a challenge, a growing number of studies have shown significant advantages of using machine learning (ML) modeling techniques with remote sensing data, especially weather radar data, for high-resolution rainfall nowcasting. To improve ML model performance, it is essential first and foremost to quantify the importance of radar variables and identify pertinent predictors of rainfall that can also be associated with domain knowledge. In this study, a set of MCS types consisting of convective cell (CC), mesoscale CC, diagonal squall line (SLD), and parallel squall line (SLP), was adopted to categorize MCS storm cells, following the fuzzy logic algorithm for storm tracking (FAST), over the Korean Peninsula. The relationships between rain rates and over 15 variables derived from data products of dual-polarimetric weather radar were investigated and quantified via five ML regression methods and a permutation importance algorithm. As an applicational example, ML classification models were also developed to predict locations of storm cells. Recalibrated ML regression models with identified pertinent predictors were coupled with the ML classification models to provide early warnings of heavy rainfall. Results imply that future work needs to consider MCS type information to improve ML modeling for nowcasting and early warning of heavy rainfall.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Toward a Globally-Applicable Uncertainty Quantification Framework for
           Satellite Multisensor Precipitation Products Based on GPM DPR

    • Free pre-print version: Loading...

      Authors: Zhe Li;Daniel B. Wright;Samantha H. Hartke;Dalia B. Kirschbaum;Sana Khan;Viviana Maggioni;Pierre-Emmanuel Kirstetter;
      Pages: 1 - 15
      Abstract: The usefulness of satellite multisensor precipitation products such as NASA’s 30-min, 0.1° Integrated Multi-satellitE Retrievals for the Global Precipitation Measurement (GPM) mission (IMERG) is hindered by their associated errors. Reliable estimates of uncertainty would mitigate this limitation, especially in near-real time when gauge observations are not available. However, creating such estimates is challenging, due to both the complicated nature of satellite precipitation errors and the lack of “ground-truth” data precisely in the places—including oceans, complex terrain, and developing countries—that could benefit most from satellite precipitation estimates. In this work, we use the GPM dual-frequency precipitation radar (DPR)-derived swath-based precipitation products as an alternative to ground-based observations to facilitate IMERG uncertainty estimation. We compare the suitability of two DPR-derived precipitation products, 2ADPR and 2BCMB, against higher fidelity ground validation multiradar multisensor (GV-MRMS) ground reference data over the contiguous United States. The 2BCMB is selected to train error models based on censored shifted gamma distribution (CSGD; a mixed discrete-continuous probability distribution). Uncertainty estimates from these models are compared against alternative error models trained on GV-MRMS. Using information from NASA’s Modern-Era Retrospective analysis for Research and Applications, version 2 (MERRA-2) reanalysis, we also demonstrate how IMERG uncertainty estimates can be further constrained using additional precipitation-related predictors. Though several critical issues remain unresolved, the proposed method shows promise for yielding robust uncertainty estimates in near-real time for IMERG and other similar precipitation products at their native resolution across the entire globe.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Machine-Learned Cloud Classes From Satellite Data for Process-Oriented
           Climate Model Evaluation

    • Free pre-print version: Loading...

      Authors: Arndt Kaps;Axel Lauer;Gustau Camps-Valls;Pierre Gentine;Luis Gómez-Chova;Veronika Eyring;
      Pages: 1 - 15
      Abstract: Clouds play a key role in regulating climate change but are difficult to simulate within Earth system models (ESMs). Improving the representation of clouds is one of the key tasks toward more robust climate change projections. This study introduces a new machine-learning-based framework relying on satellite observations to improve understanding of the representation of clouds and their relevant processes in climate models. The proposed method is capable of assigning distributions of established cloud types to coarse data. It facilitates a more objective evaluation of clouds in ESMs and improves the consistency of cloud process analysis. The method is built on satellite data from the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument labeled by deep neural networks with cloud types defined by the World Meteorological Organization (WMO), using cloud-type labels from CloudSat as ground truth. The method is applicable to datasets with information about physical cloud variables comparable to MODIS satellite data and at sufficiently high temporal resolution. We apply the method to alternative satellite data from the Cloud_cci project (ESA Climate Change Initiative), coarse-grained to typical resolutions of climate models. The resulting cloud-type distributions are physically consistent and the horizontal resolutions typical of ESMs are sufficient to apply our method. We recommend outputting crucial variables required by our method for future ESM data evaluation. This will enable the use of labeled satellite data for a more systematic evaluation of clouds in climate models.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Spectral Energy Model-Driven Inversion of XCO2 in IPDA Lidar
           Remote Sensing

    • Free pre-print version: Loading...

      Authors: Haowei Zhang;Ge Han;Xin Ma;Siwei Li;Hao Xu;Tianqi Shi;Jianye Yuan;Wanqin Zhong;Yanran Peng;Jingjing Xu;Wei Gong;
      Pages: 1 - 9
      Abstract: Carbon observation satellites based on passive theory (e.g., OCO-2/3, GOSAT-1/2, and TanSat) have relatively high carbon dioxide column concentration (XCO2) accuracy when the observation conditions are met. Passive satellites have data bias and coverage deficiencies due to cloud cover, low albedo, low-light conditions, and aerosol scattering, resulting in carbon observation satellites based on passive theory that cannot meet the demand for high-precision, all-day, all-weather XCO2 monitoring. Active detection satellites are urgently needed to support global carbon sources, sinks, and carbon neutrality. China intends to launch a sensor satellite with active detection of XCO2 in the coming years. In this work, based on the satellite’s scaled-down airborne experiments, a spectral energy model was developed to optimize the conventional inversion algorithm and achieve a more accurate XCO2 inversion. The 1.572- $mu text{m}$ integrated path differential absorption (IPDA) lidar column length is used indirectly to evaluate the accuracy of the spectral energy model for signal extraction. Also, the experimental results show that the accuracy of the signal extracted by the 1.572- $mu text{m}$ IPDA lidar column length is 0.74 and 6.20 m at sea and on land based on the indirect evaluation of the length of the 1.572- $mu text{m}$ IPDA lidar column length. The optimized XCO2 was evaluated (standard deviation as an evaluation metric) and its XCO2 standard deviation reduced by 31%, 63%, and 66% in the ocean, plains, and mountains, respectively. Our algorithm can obtain the XCO2 with a consistent trend by using XCO2 from the OCO-2 satellite as a reference. The calculated XCO2 is more accurat- in areas dominated by anthropogenic factors (plains), due to the accuracy of the IPDA detection mechanism. This algorithm improves the accuracy and robustness of XCO2 inversion and has important reference significance for the IPDA lidar carried by China’s satellites to be launched in this year.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Tomographic Reconstruction of Water Vapor Density Fields From the
           Integration of GNSS Observations and Fengyun-4A Products

    • Free pre-print version: Loading...

      Authors: Biyan Chen;Jingshu Tan;Wei Wang;Wujiao Dai;Minsi Ao;Chunhua Chen;
      Pages: 1 - 12
      Abstract: The potential of precipitable water vapor (PWV) maps retrieved by remote sensing satellites can address the geometry defect of global navigation satellite system (GNSS) observations in tropospheric tomography. The second-generation geostationary meteorological satellite Fengyun-4A (FY-4A) of China can provide PWV products with high spatial (4 km) and temporal (15 min) resolutions. This article presents the first study on water vapor tomography by integrating GNSS measurements and FY-4A products using the node-based parameterized method. Layer PWV (LPW) products of FY-4A instead of the total PWV are adopted, which can increase the rank of the tomographic equation significantly. The integrated tomography model is validated with observational data collected over the three-month period of June to August 2020 from 124 GNSS stations in Hunan province, China. Assessments using radiosonde and European Centre for Medium-Range Weather Forecasts ReAnalysis 5 (ERA5) data demonstrate the better performance of the integrated model against the traditional model using GNSS data alone. In the assessment with radiosonde profiles, the integrated model improves the tomographic solutions upon the traditional model by 40.58% and 36.33% for 30- and 15-min resolutions, respectively. Root mean square errors (RMSEs) of density differences between ERA5 and the integrated model vary from 1.24 to 2.82 g/ $text{m}^{3}$ throughout the study area. RMSEs vertically decrease from $sim 5$ g/ $text{m}^{3}$ at the bottom to $sim 0.5$ g/ $text{m}^{3}$ at the top layer of about 10 km. This work demonstrate- the benefit of high-quality FY-4A products to GNSS tomography because they can effectively mitigate the ill-posed problem of inverse process.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Improved Aerosol Lidar Ratio Profile by Introducing Pseudo-Constant

    • Free pre-print version: Loading...

      Authors: Hongzhu Ji;Siying Chen;Pan Guo;Jingxi He;Yuefeng Zhao;Xinye Fan;
      Pages: 1 - 10
      Abstract: Aerosol lidar ratio (LR) is important for the inversion of aerosol optical characteristics. Single aerosol LR value adopted in the inversion not only causes difficulties in improving aerosol inversion accuracy but also hinders the analysis of aerosol types. In this article, a pseudo-constant method with certain universality is proposed for inverting aerosol LR profiles. Simulation results show a decrease in the mean absolute error (MAE) of the aerosol LR from 3.57% (iteration method) and 9.96% (Fernald method) to 1.82% (pseudo-constant method). In the near-surface region (< 1.5 km), the MAE of aerosol LR decreases from 4.97% (iteration) and 21.66% (Fernald) to 1.54% (pseudo-constant). Moreover, the inversion accuracy of aerosol LR obtained from pseudo-constant method is less influenced by the precision of the pre-estimated aerosol LR value than that from iteration method. Experimentally, the aerosol LR and aerosol extinction coefficient (AEC) profiles obtained from the pseudo-constant method are more accurate than those from the iteration and Fernald methods. In addition, both the simulations and experiments indicate that the pseudo-constant method has higher efficiency; the processing time is shortened from approximately 5 s/signal (iteration) to approximately 0.1 s/signal. Finally, aerosol types can be distinguished to a certain degree by combining the improved aerosol LR profile.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Fast Piecewise-Defined Neural Network Method to Retrieve Temperature and
           Humidity Profile for the Vertical Atmospheric Sounding System of
           FengYun-3E Satellite

    • Free pre-print version: Loading...

      Authors: Wenguang Bai;Peng Zhang;Hui Liu;Wenjian Zhang;Chengli Qi;Gang Ma;Guannan Li;
      Pages: 1 - 10
      Abstract: A fast piecewise-defined neural network (PDNN) method is presented to produce accurate atmospheric temperature and humidity profiles from satellite hyperspectral infrared (IR) and microwave (MW) observations in all-sky conditions. The PDNN method relies on a novel classification approach and a principal component-based neural network (NN) function to better capture the nonlinear relationship between the spectral radiances and atmospheric state vectors. The algorithm was designed to only rely on satellite measurements. Large datasets were used for network training to make the retrieval robust to random errors in the reference data. For each retrieved profile, an effective quality indicator (QI) was obtained by training against the absolute value of the retrieval error. In addition, a reliable rain cloud flag was generated based on the scattering difference of cloud particles between the 50 and 118 GHz channels. Besides the independent reanalysis of field data, the algorithm’s performance was also evaluated using radiosonde measurements. Preliminary validation shows that the best temperature retrieval occurred in the mid-troposphere (around 1.0 K). Depending on the reference data, errors in the boundary layer typically ranged from 1.8 to 2.5 K. For water vapor, the retrieval error was less than 22% in the low-troposphere and less than 35% in the mid- and upper-troposphere when validated against the reanalysis field. The humidity error was approximately 10% lower when compared to radiosondes. The PDNN is used operationally to produce atmospheric temperature and moisture soundings from the Vertical Atmospheric Sounding System (VASS) of the FengYun-3E (FY-3E) satellite, an early-morning-orbit meteorological satellite launched in 2021.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Estimating Near-Surface Concentrations of Major Air Pollutants From Space:
           A Universal Estimation Framework LAPSO

    • Free pre-print version: Loading...

      Authors: Songyan Zhu;Jian Xu;Meng Fan;Chao Yu;Husi Letu;Qiaolin Zeng;Hao Zhu;Hongmei Wang;Yapeng Wang;Jiancheng Shi;
      Pages: 1 - 11
      Abstract: Like many other countries, China is still facing severe air pollution issues after extensive efforts. The difficulties in deriving near-surface concentrations from satellite measurements restrict the application of remote sensing of large-scale surface air quality. Aiming at providing daily accurate near-surface ail pollution estimates (PM2.5, PM10, O3, NO2, SO2, and CO), we propose a robust estimation framework called learning air pollutants from satellite observations (LAPSO). The principle of LAPSO is to derive a nonlinear relationship between surface pollutant concentrations of interest and satellite observations with the aid of meteorological reanalyzes based on deep learning techniques. The LAPSO framework is superior to other algorithms due to its robust retrieval performance, independence from chemical transport models (CTMs), lower hardware requirements, and a user-friendly interface. The retrieval results of LAPSO were in good agreement with ground-level measurements according to extensive cross-validation at 1628 sites ( $text{R}^{2}>$ 0.8 in polluted areas and uncertainty $ll 5~mu text{g}/text{m}^{3}$ for most pollutants) in China. The framework also showed a strong capability to capture the temporal variability of different air pollutants. By comparing with the estimation results from different satellite platforms, TROPOspheric monitoring instrument (TROPOMI) onboard the Sentinel-5P demonstrated marginally better performance for estimating PM2.5. Although the selection of satellite observations did not significantly affect the results of O3 estimation, the number and spatial sampling density of in situ sites imposed large impacts on O3 estimation performance. The success of LAPSO for estimating near-surface concentrations from satellite remote sensing at an enhanced spatiotemporal resolutio- is expected to serve the continuous and dynamical monitoring of regional and global air pollution.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Improving the Accuracy of MODIS Near-Infrared Water Vapor Product Under
           all Weather Conditions Based on Machine Learning Considering Multiple
           Dependence Parameters

    • Free pre-print version: Loading...

      Authors: Jiafei Xu;Zhizhao Liu;
      Pages: 1 - 15
      Abstract: We developed six machine-learning-based calibration models to improve the all-weather accuracy of precipitable water vapor (PWV) product from near-infrared (NIR) observations of the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument, i.e., MOD05 PWV. The six machine learning approaches are: back propagation neural network (BPNN), gradient boosting decision tree (GBDT), generalized regression neural network (GRNN), K-nearest neighbor (KNN), multilayer perceptron neural network (MLPNN), and extreme gradient boosting (XGBoost). The input of the models included MOD05 PWV, latitude, longitude, elevation, cloud, season, and solar zenith angle, in association with the quality of the MOD05 PWV product. PWV data measured from in situ 453 Global Positioning System (GPS) stations in Australia in 2017 were utilized as the target water vapor data for model training. The validation results in Australia in 2018–2019 indicate that the models can significantly improve the all-weather quality [increased $R^{2}$ , reduced root-mean-square error (RMSE), and reduced mean bias (MB)] of the MOD05 PWV product, exhibiting a reduction in RMSE of 53.33% for BPNN, 55.25% for GBDT, 53.24% for GRNN, 37.81% for KNN, 54.98% for MLPNN, and 55.16% for XGBoost. The $R^{2}$ was 0.58–0.79 and the MB was 0.12–0.72 mm, much better than the MOD05 all-weather PWV product ( $R^{2} = 0.26$ and MB = −1.75 mm). Different from previous studies that focused on clear-sky conditions only, this work is the first one to enhance the quality of official MODIS NIR PWV products under all-weather conditions, reducing the impact of clouds on PWV products.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A New Cloud and Haze Mask Algorithm From Radiative Transfer Simulations
           Coupled With Machine Learning

    • Free pre-print version: Loading...

      Authors: Yingzi Jiao;Ming Zhang;Lunche Wang;Wenmin Qin;
      Pages: 1 - 16
      Abstract: Mainstream satellite cloud masking algorithms are prone to mis-masking in haze-polluted areas, which may cause errors in aerosol radiative effect calculations and attribution of surface solar radiance changes; thereby, distinguishing between clouds and haze is critical to obtaining accurate land and atmospheric data products. Existing cloud and haze mask algorithms based on the threshold method may require us to spend a lot of manpower to perform multiple threshold tests; in addition, the obtained thresholds are only applicable to particular sensors, which limits the generality of the threshold-based cloud and haze mask algorithms. In this study, a new cloud and haze mask algorithm based on a combination of radiative transfer simulations and machine learning text simulation-based cloud and haze masking (SCHM) is proposed and applied to MODIS images. When we simulated the apparent reflectance of the first seven visible and text near-infrared channels of MODIS, the CALIOP and AERONET data verification results showed that the SCHM algorithm achieved 85.16% and 90.08% hit rates for cloud and haze recognition, respectively. When we added three thermal infrared channels (20, 31, and 35 bands) for simulation, the cloud and haze hit rates were improved to approximately 85.72% and 90.62%, respectively. This indicates that the SCHM algorithm can improve the accuracy of detection results by improving the radiative transfer simulation parameters. Compared with existing threshold-based methods, the SCHM algorithm has the advantages of simple logic, convenient modification, and flexible configuration.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Modeling of Rain Drop Size Distribution in Association With Convective and
           Cloud Parameter Over a Tropical Location

    • Free pre-print version: Loading...

      Authors: Arijit De;Animesh Maitra;
      Pages: 1 - 9
      Abstract: The difference in rain drop size distribution (DSD) characteristics between convective and stratiform rain and in monsoon and premonsoon season have been investigated over a tropical location, Kolkata, using seven years of data collected from a Joss–Waldvogel disdrometer (JWD). During the Indian summer monsoon (ISM) and the premonsoon season, Kolkata gets substantial rain. Interannual variability of DSD has been evident. The normalized drop size distributions of DSD show a double peak at 20–50 mm/h rain rate due to drop break up and coalescence process. For convective rain, the concentration of higher drops ( $>$ 3 mm) is more significant compared to stratiform rain. In the premonsoon period, a significant contribution of normalized DSD in the afternoon to late evening has been evident due to nor’westers. The contribution of normalized DSD of higher drops ( $>$ 3 mm) is more in the premonsoon convective rain compared to monsoon convective rain. The larger slope of $mu $ – $Lambda $ relationship has been observed in the premonsoon convective period which corresponds to higher $D_{m}$ values. The relationship between radar reflectivity ( $Z$ ) and rain rate ( $R$ ) distinguishes the monsoon and premonsoon seasons. Radiometric measurements have been used to examine the effects of the atmospheric instability parameter convective available potential energy (CAPE) and cloud-base height (CBH) on the formation of larger drop sizes. This study will be useful for understanding precipitation microphysics and improving the quantitative precipitation estimation (QPE) algorithm.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Responses of GNSS ZTD Variations to ENSO Events and Prediction Model Based
           on FFT-LSTME

    • Free pre-print version: Loading...

      Authors: Tengli Yu;Ershen Wang;Shuanggen Jin;Yong Wang;Jing Huang;Xiao Liu;Wei Zhan;
      Pages: 1 - 17
      Abstract: The El Niño-Southern Oscillation (ENSO) event often causes natural disasters in mainland China. Existing quantitative analysis of ENSO event’s effects on climate change in mainland China is insufficient. The monthly scale prediction effectiveness of ENSO events is still low. Global Navigation Satellite System (GNSS) can estimate zenith tropospheric delay (ZTD) with high accuracy, which can study ZTD responses to ENSO and improve the prediction accuracy of ENSO events. This study quantitatively analyzed the response patterns of GNSS ZTD time–frequency variation to ENSO events in mainland China. The monthly multivariate ENSO index (MEI) thresholds for GNSS ZTD anomaly response to ENSO events are (−1.12, 1.92) for the tropical monsoon zone (TPMZ), (−1.12, 1.61) for the subtropical monsoon zone (SMZ), (−1.19, 1.62) for the temperate monsoon zone (TMZ), (−1.26, 1.64) for the temperate continental zone (TCZ), and (−1.22, 1.72) for the mountain plateau zone (MPZ). The ENSO event causes the amplitude of the nine-month variation period to decrease and the amplitude of the 0.8–3-month period to increase for the GNSS ZTD in mainland China. Furthermore, a forecasting model is proposed by integrating fast Fourier transform and long short-term memory extended (FFT-LSTME). The model uses monthly MEI as the primary input and the GNSS ZTD reconstruction sequence that responds to ENSO as the auxiliary input. It can predict ENSO events in the next 24 months with an index of agreement (IA) of 91.56% and a root mean square error (RMSE) of 0.25. The RMSE is optimized by 70.48%, 43.95%, and 11.6% when compared with radial basis function (RBF), LSTM, and FFT-LSTM.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Assessment on the Diurnal Cycle of Cloud Covers of Fengyun-4A
           Geostationary Satellite Based on the Manual Observation Data in China

    • Free pre-print version: Loading...

      Authors: Yongen Liang;Min Min;Yu Yu;Wang Xi;Pan Xia;
      Pages: 1 - 18
      Abstract: Complicated and regionally representative diurnal cycle characteristics of clouds may introduce some errors in the cloud mask (CLM) algorithm of the Geostationary (GEO) meteorological satellite imaging system, which are very difficult to be assessed by using analogous products of fixed-passing polar-orbiting satellites. In this investigation, the diurnal cycle of the performance of the CLM algorithm of the Advanced Geosynchronous Radiation Imager onboard the China Fengyun-4A satellite (FY-4A/AGRI) is validated by using manually observed cloud covers (CC) at 25 ground-based stations in China. The results indicate that the CCs calculated by the FY-4A/AGRI CLM algorithm are overestimated at 11:00 BJT (Beijing Time) and 14:00 BJT (around noon) and underestimated at 08:00 BJT and 20:00 BJT (in the morning and evening) at most stations. In summer, compared with other seasons, the CCs obtained from the FY-4A/AGRI over northern China and the Tibetan Plateau are much better, consistent with the manual observations, but the situation is the opposite in southern China. The CC results retrieved at the vegetation surface by FY-4A/AGRI, however, show the best and stable performance. Because of that, the two independent cloud tests induce most of the overestimations, and some sensitivity experiments for the CLM algorithm are conducted. The results show that the best improvement effect is achieved after only closing one cloud test using the $3.8mu text{m}$ band. Many extremely overestimated CC samples (about 56.3%) are eliminated. After that, the FY-4A/AGRI CLM product is more reasonable compared with the corresponding infrared and visible imageries.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • The Status and Influencing Factors of Surface Water Dynamics on the
           Qinghai-Tibet Plateau During 2000–2020

    • Free pre-print version: Loading...

      Authors: Qinwei Ran;Filipe Aires;Philippe Ciais;Chunjing Qiu;Ronghai Hu;Zheng Fu;Kai Xue;Yanfen Wang;
      Pages: 1 - 14
      Abstract: The Qinghai–Tibet Plateau is rich in water resources with numerous lakes, rivers, and glaciers, and, as a source of many rivers in Central Asia, it is known as the Asian Water Tower. Under global climate change, it is critical to understand the current influencing factors on surface water area in this region. Although there are numerous studies on surface water mapping, they are still limited by temporal/spatial resolution and record length. Moreover, the complicated topographic condition makes it challenging to map the surface water accurately. Here, we proposed an automatic two-step annual surface water classification framework using long time-series Landsat images and topographic information based on the Google Earth Engine (GEE) platform. The results showed that the producer accuracy (PA) and user accuracy (UA) of the surface water map in the Qinghai–Tibet Plateau in 2020 were 99% and 90%, respectively, and the Kappa coefficient reached 0.87. Our dataset showed high consistency with high-resolution images, indicating that the proposed large-scale water mapping method has great application potential. Furthermore, a new annual surface water area dataset on the Qinghai–Tibet Plateau from 2000 to 2020 was generated, and its relationship with climate, vegetation, permafrost, and glacier factors was explored. We found that the mean surface water area was about 59 481 km2, and there was a significant increasing trend (=322 km2/year, $p < 0.01$ ) during 2000–2020 in the plateau. Greening, warming, and wetting climate conditions contributed to the increase of surface water area. Active layer thickness and permafrost types may be the most related to the decrease of surface water area. This study provides important information for ecological assessment and-protection of the plateau and promotes the implementation of sustainable development goals related to surface water resources.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • High Latitude Sea Surface Skin Temperatures Derived From Saildrone
           Infrared Measurements

    • Free pre-print version: Loading...

      Authors: Chong Jia;Peter J. Minnett;Malgorzata Szczodrak;Miguel Izaguirre;
      Pages: 1 - 14
      Abstract: From May 15 to October 11, 2019, six Saildrone uncrewed surface vehicles (USVs) were deployed for 150-day cruises collecting a suite of atmospheric and oceanographic measurements from Dutch Harbor, Alaska, transiting the Bering Strait into the Chukchi Sea and the Arctic Ocean. Two Saildrones funded by the National Aeronautics and Space Administration (NASA), SD-1036 and SD-1037, were equipped with infrared (IR) radiation pyrometers in a “unicorn” structure on the deck for the determination of the ocean sea surface skin temperature (SST $_{mathrm {skin}}$ ). We present an algorithm to derive SSTskin from the downward- and upward-looking radiometers and estimate the main contributions to the inaccuracy of SSTskin. After stringent quality control of data and eliminating measurements influenced by sea ice and precipitation, and restricting the acceptable tilt angles of the USV based on radiative transfer simulations, SSTskin can be derived to an accuracy of approximately 0.12 K. The error budget of the derived SSTskin is developed, and the largest component comes from the instrumental uncertainties, assuming that the viewing geometry is adequately determined. Thus, Saildrones equipped with these sensors could provide sufficiently accurate SSTskin retrievals for studying the physics of the thermal skin effect, in conjunction with accurate subsurface thermometer measurements, and for validating satellite-derived SSTskin fields at high latitudes.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Impact of Rain Effects on L-Band Passive Microwave Satellite Observations
           Over the Ocean

    • Free pre-print version: Loading...

      Authors: Xuchen Jin;Xianqiang He;Difeng Wang;Jianyun Ying;Fang Gong;Qiankun Zhu;Chenghu Zhou;Delu Pan;
      Pages: 1 - 16
      Abstract: L-band passive microwave remote sensing of ocean surfaces is hampered by uncertainties due to the contribution of precipitation. However, modeling and correcting rain effect are complicated because all the precipitation contributions from the atmosphere and sea surface, including the rain effect in the atmosphere, water refreshing, rain-perturbed sea surface, and rain-induced local wind, are coupled. These rain effects significantly alter L-band microwave satellite measurements. This study investigates the impact of precipitation on the satellite measured brightness temperature (TB) and proposes a correction method. The results show that the rain-induced TB increase is approximately 0.5–1.4 K in the atmosphere, depending on the incidence angle and polarization. Moreover, the rain effect on sea surface emissions is more significant than that in the atmosphere. The result shows that rain effects on sea surface emissions are higher than 3 K for both polarizations when the rain rate is higher than 20 mm/h. We validate the rain effect correction model based on Soil Moisture Active Passive (SMAP) observations. The results show that the TBs at the top of the atmosphere (TOA) simulated by the model are in a good agreement with the SMAP observations, with root-mean-square errors (RMSEs) of 1.137 and 1.519 K for the horizontal and vertical polarizations, respectively, indicating a relatively high accuracy of the established model. Then, a correction model is applied to sea surface salinity (SSS) retrieval for analysis, and the results show that the developed model corrects the underestimation in SSS retrieval. Finally, the rain effect correction model is validated with Aquarius observations in three regions, and it is found that the RMSEs of the corrected TOA TBs range from 1.013 to 1.608 K, which is higher than those without rain effect correction (RMSEs range from 2.078 to 3.894 K). Overall, the model developed in this study provides relatively a good ac-uracy for rain effect correction.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Significant Wave Height Retrieval Based on Multivariable Regression Models
           Developed With CYGNSS Data

    • Free pre-print version: Loading...

      Authors: Changyang Wang;Kegen Yu;Kefei Zhang;Jinwei Bu;Fangyu Qu;
      Pages: 1 - 15
      Abstract: This study utilizes L1B level data from reflected global navigation satellite system (GNSS) signals from the Cyclone GNSS (CYGNSS) mission to estimate sea surface significant wave height (SWH). The normalized bistatic radar cross Section (NBRCS), the leading edge slope (LES), the signal-to-noise ratio (SNR), and the delay-Doppler map average (DDMA) are used as the key variables for the SWH retrieval. Eight other parameters, including instrument gain and scatter area, are also utilized as auxiliary variables to enhance the SWH retrieval performance. A variety of multivariable regression models are investigated to clarify the relationship between the SWH and the variables by using the following five methods: stepwise linear regression, Gaussian support vector machine, artificial neural network, sparrow search algorithm–extreme learning machine, and bagging tree (BT). Results show that, among the five regression models developed, the BT model performs the best with the root mean square error (RMSE) of 0.48 m and the correlation coefficient (CC) of 0.82 when testing one million sets of data randomly selected, while the RMSE and CC of BT model are 0.44 m and 0.73 in the 4500 National Data Buoy Center (NDBC) buoy testing dataset. Meanwhile, the BT model also has the best generalization ability, which means that it performs well in practical applications. In addition, the impacts of different input variables, the size of the training dataset, and the sea surface wind speed are also investigated. These findings are anticipated to serve as helpful guides for creating future SWH retrieval algorithms that are more advanced.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Reconstruction of 3-D Ocean Chlorophyll a Structure in the Northern Indian
           Ocean Using Satellite and BGC-Argo Data

    • Free pre-print version: Loading...

      Authors: Qiwei Hu;Xiaoyan Chen;Yan Bai;Xianqiang He;Teng Li;Delu Pan;
      Pages: 1 - 13
      Abstract: We present a novel method using satellite and biogeochemical Argo (BGC-Argo) data to retrieve the 3-D structure of chlorophyll $a$ (Chla) in the northern Indian Ocean (NIO). The random forest (RF)-based method infers the vertical distribution of Chla using the near-surface and vertical features. The input variables can be divided into three categories: 1) near-surface features acquired by satellite products; 2) vertical physical properties obtained from temperature and salinity profiles collected by BGC-Argo floats; and 3) the temporal and spatial features, i.e., day of the year, longitude, and latitude. The RF-model is trained and evaluated using a large database including 9738 profiles of Chla and temperature-salinity properties measured by BGC-Argo floats from 2011 to 2021, with synchronous satellite-derived products. The retrieved Chla values and the validation dataset (including 1948 Chla profiles) agree fairly well, with $R^{2} = 0.962$ , root-mean-square error (RMSE) = 0.012, and mean absolute percent difference (MAPD) = 11.31%. The vertical Chla profile in the NIO retrieved from the RF-model is more accurate and robust compared to the operational Chla profile datasets derived from the neural network and numerical modeling. A major application of the RF-retrieved Chla profiles is to obtain the 3-D Chla structure with high vertical resolution. This will help to quantify phytoplankton productivity and carbon fluxes in the NIO more accurately. We expect that RF-model can be used to develop long-time series products to understand the variability of 3-D Chla in future climate change scenarios.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • 2-D Magnetic Resonance Tomography With an Inaccurately Known Larmor
           Frequency Based on Frequency Cycling

    • Free pre-print version: Loading...

      Authors: Jiannan Liu;Baofeng Tian;Chuandong Jiang;Ruixin Miao;Yanju Ji;
      Pages: 1 - 13
      Abstract: When using magnetic resonance tomography (MRT) for imaging 2-D or 3-D water-bearing structures in a subsurface, the transmitting frequency must be the same as the Larmor frequency. Due to the inhomogeneity and noise interference in a geomagnetic field, it is difficult to determine the precise Larmor frequency using a magnetometer, resulting in unknown frequency offsets and inaccurate estimations of water content and relaxation time ( $T_{2}^{*}$ ). To solve the 2-D MRT imaging problem in the case of an unknown frequency offset, a frequency cycling method is proposed in this article. This method takes the estimated Larmor frequency as the center, uses two frequencies with the same offset for transmitting, then combines the acquired MRT signals to obtain frequency-cycled data, and finally uses the off-resonance kernel function for inversion. Based on MRT forward modeling and QT inversion, we conduct synthetic data experiments on a complex model with three water-bearing structures and test the 2-D imaging results of the frequency-cycled data. The results show that the water content and $T_{2}^{*}$ distribution obtained by the inversion of the frequency-cycled data can accurately reflect the water-bearing structure, which is better than the results of the assumed on-resonance case. In addition, the phase correction method presented in this article significantly improves the accuracy of 2-D MRT estimated aquifer properties under low resistivity conditions. Finally, the validity and accuracy of the frequency cycling method are verified by comparing the inversion results of the data with known drilling data collected in field measurements.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Per-Pixel Uncertainty Quantification and Reporting for Satellite-Derived
           Chlorophyll-a Estimates via Mixture Density Networks

    • Free pre-print version: Loading...

      Authors: Arun M. Saranathan;Brandon Smith;Nima Pahlevan;
      Pages: 1 - 18
      Abstract: Mixture density networks (MDNs) have emerged as a powerful tool for estimating water-quality indicators, such as chlorophyll-a (Chl $a$ ) from multispectral imagery. This study validates the use of an uncertainty metric calculated directly from Chl $a$ estimates of the MDNs. We consider multispectral remote sensing reflectance spectra ( $R_{text {rs}}$ ) for three satellite sensors commonly used in aquatic remote sensing, namely, the ocean and land colour instrument (OLCI), multispectral instrument (MSI), and operational land imager (OLI). First, a study on a labeled database of colocated in situ Chl $a$ and $R_{text {rs}}$ measurements clearly illustrates that the suggested uncertainty metric accurately captures the reduced confidence associated with test data, which is drawn for a different distribution than the training data. This change in distribution maybe due to: 1) random noise; 2) uncertainties in the atmospheric correction; and 3) novel (unseen) data. The experiments on the labeled in situ dataset show that the estimated uncertainty has a correlation with the expected predictive error and can be used as a bound on the predictive error for most samples. To illustrate the ability of the MDNs in generating consistent products from multiple sensors, per-pixel uncertainty maps for three near-coincident images of OLCI, MSI, and OLI are produced. The study also examines temporal trends in OLCI-derived Chl $a$ and the associated uncertainties at selected locations over a calendar year. Future work will include unce-tainty estimation from MDNs with a multiparameter retrieval capability for hyperspectral and multispectral imagery.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A New Spectral Harmonization Algorithm for Landsat-8 and Sentinel-2 Remote
           Sensing Reflectance Products Using Machine Learning: A Case Study for the
           Barents Sea (European Arctic)

    • Free pre-print version: Loading...

      Authors: Muhammad Asim;Atsushi Matsuoka;Pål Gunnar Ellingsen;Camilla Brekke;Torbjørn Eltoft;Katalin Blix;
      Pages: 1 - 19
      Abstract: The synergistic use of Landsat-8 operational land imager (OLI) and Sentinel-2 multispectral instrument (MSI) data products provides an excellent opportunity to monitor the dynamics of aquatic ecosystems. However, the merging of data products from multisensors is often adversely affected by the difference in their spectral characteristics. In addition, the errors in the atmospheric correction (AC) methods further increase the inconsistencies in downstream products. This work proposes an improved spectral harmonization method for OLI and MSI-derived remote sensing reflectance ( ${R_{rs}}$ ) products, which significantly reduces uncertainties compared to those in the literature. We compared ${R_{rs}}$ retrieved via state-of-the-art AC processors, i.e., Acolite, C2RCC, and Polymer, against ship-based in situ ${R_{rs}}$ observations obtained from the Barents Sea waters, including a wide range of optical properties. Results suggest that the Acolite-derived ${R_{rs}}$ has a minimum bias for our study area with median absolute percentage difference (MAPD) varying from 9% to 25% in the blue–green bands. To spectrally merge OLI and MSI, we develop and apply a new machine learning-based bandpass adjustment (BA) model to near-simultaneous OLI and MSI images acquired in the years from 2018 to 2020. Compared to a conventional linear adjustment, we demonstrate that the spectral difference is significantly reduced from $sim 6$ % to 12% to $sim 2$ %- to ${< } {10%}$ in the common OLI-MSI bands using the proposed BA model. The findings of this study are useful for the combined use of OLI and MSI ${R_{rs}}$ products for water quality monitoring applications. The proposed method has the potential to be applied to other waters.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Deep Multi-Look Sequence Processing for Synthetic Aperture Sonar Image
           Segmentation

    • Free pre-print version: Loading...

      Authors: Isaac D. Gerg;Vishal Monga;
      Pages: 1 - 15
      Abstract: Deep learning has enabled significant improvements in semantic image segmentation, especially in underwater imaging domains such as side scan sonar (SSS). In this work, we apply deep learning to synthetic aperture sonar (SAS) imagery, which has an advantage over traditional SSS in which SAS produces coherent high- and constant-resolution imagery. Despite the successes of deep learning, one drawback is the need for abundant labeled training data to enable success. Such abundant labeled data are not always available as in the case of SAS where collections are expensive and obtaining quality ground-truth labels may require diver intervention. To overcome these challenges, we propose a domain-specific deep learning network architecture utilizing a unique property to complex-valued SAS imagery: the ability to resolve angle-of-arrival (AoA) of acoustic returns through $k$ -space processing. By sweeping through consecutive incrementally advanced AoA bandpass filters (a process sometimes referred to as multi-look processing), this technique generates a sequence of images emphasizing angle-dependent seafloor scattering and motion from biologics along the seafloor or in the water column. Our proposal, which we call multi-look sequence processing network (MLSP-Net), is a domain-enriched deep neural network architecture that models the multi-look image sequence using a recurrent neural network (RNN) to extract robust features suitable for semantic segmentation of the seafloor without the need for abundant training data. Unlike previous segmentation works in SAS, our model ingests a complex-valued SAS image and affords the ability to learn the AoA filters in $k$ -space as part of the training procedure. We show the results on a challenging real-world SAS database, and despite the lack of abundant training data, our proposed method shows-superior results over state-of-the-art techniques.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • On the Use of Dual Co-Polarized Radar Data to Derive a Sea Surface Doppler
           Model—Part 1: Approach

    • Free pre-print version: Loading...

      Authors: Vladimir Kudryavtsev;Shengren Fan;Biao Zhang;Bertrand Chapron;Johnny A. Johannessen;Artem Moiseev;
      Pages: 1 - 13
      Abstract: This article proposes a Doppler velocity (DV) model based on dual co-polarized (co-pol) decomposition of a normalized radar cross section of an ocean surface on polarized Bragg scattering and nonpolarized (NP) radar returns from breaking wave components. The dual co-pol decomposition provides a quantitative description of resonant and NP scattering, as well as their dependence on the incident angle, azimuth, and wind speed. Subsequently, the contributions of the facet (resonant Bragg waves and breakers) velocities, tilt, and hydrodynamic modulations due to long waves to the resulting DV can be quantified. The tilt modulation contributions to DV are estimated using the measured/empirical tilt modulation transfer function (MTF). The hydrodynamic modulations are mostly dominated by wave breaking and are estimated using a semiempirical model based on in situ measurements. In addition to the VV and HH radar data, which are required for dual co-pol decomposition and tilt MTF estimates, the surface wave spectrum is required in the DV determination for a given radar observation geometry. In this article, qualitative and quantitative consistencies are presented between the model simulations and the empirical CDOP model. In a companion paper, a DV analysis is presented to analyze the Sentinel-1 synthetic aperture radar measurements and collocated in situ measurements of surface wind and wave spectra.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Image-to-Height Domain Translation for Synthetic Aperture Sonar

    • Free pre-print version: Loading...

      Authors: Dylan Stewart;Austin Kreulach;Shawn F. Johnson;Alina Zare;
      Pages: 1 - 13
      Abstract: Synthetic aperture sonar (SAS) intensity statistics are dependent upon the sensing geometry at the time of capture. Estimating bathymetry from acoustic surveys is challenging. While several methods have been proposed to estimate seabed relief via intensity, we develop the first large-scale study that relies on deep learning models. In this work, we pose bathymetric estimation from SAS surveys as a domain translation problem of translating intensity to height. Since no dataset of coregistered seabed relief maps and sonar imagery previously existed to learn this domain translation, we produce the first large simulated dataset containing coregistered pairs of seabed relief and intensity maps from two unique sonar data simulation techniques. We apply four types of models, with varying complexity, to translate intensity imagery to seabed relief: a shape-from-shading (SFS) approach, a Gaussian Markov random field (GMRF) approach, a conditional Generative Adversarial Network (cGAN), and UNet architectures. Each model is applied to datasets containing sand ripples, rocky, mixed, and flat sea bottoms. Methods are compared in reference to the coregistered simulated datasets using L1 error. Additionally, we provide results on simulated and real SAS imagery. Our results indicate that the proposed UNet architectures outperform an SFS, a GMRF, and a pix2pix cGAN model.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Assessing the Relative Performance of GNSS-R Flood Extent Observations:
           Case Study in South Sudan

    • Free pre-print version: Loading...

      Authors: Brandi Downs;Albert J. Kettner;Bruce D. Chapman;G. Robert Brakenridge;Andrew J. O’Brien;Cinzia Zuffada;
      Pages: 1 - 13
      Abstract: Flooding is one of the deadliest and costliest natural disasters. Climate change-induced flooding events are increasing worldwide, disproportionately impacting low-income and developing communities. While early warning systems save lives, satellite-based observation systems are vital for the disaster relief and recovery phases. Current satellite-based operational flood products are largely based on either optical remote-sensing methods, which exhibit a limited ability to detect water through clouds and vegetation, or microwave remote sensing, which provides relatively low spatial and temporal resolution. New small satellite constellations using radar or GNSS reflectometry (GNSS-R) have been shown to enhance our ability to overcome these deficiencies. In this work, we quantify the performance of using GNSS-R measurements from the NASA Cyclone Global Navigation Satellite Systems (CYGNSS) satellite constellation to map surface water in South Sudan and the Sudd wetland in comparison with a set of representative operational products. We make quantitative comparisons of our results with operational flood products based on Visible Infrared Imaging Radiometer Suite (VIIRS) and MODIS and with C-band Sentinel-1 synthetic aperture radar. We find that our method detects 35.4% more surface water than Sentinel-1, while the VIIRS- and MODIS-based products underestimate by 4.8% and 83.7%, respectively. We use several metrics commonly used to evaluate classification performance: precision, true positive rate (TPR), true negative rate (TNR), F2-score, and the Matthews correlation coefficient (MCC) and assess the comparisons in this statistical framework. We discuss the consequences of our findings, including ways CYGNSS data may enhance current flood products and assist decision-makers and emergency managers.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • DGNet: Distribution Guided Efficient Learning for Oil Spill Image
           Segmentation

    • Free pre-print version: Loading...

      Authors: Fang Chen;Heiko Balzter;Feixiang Zhou;Peng Ren;Huiyu Zhou;
      Pages: 1 - 17
      Abstract: Successful implementation of oil spill segmentation in synthetic aperture radar (SAR) images is vital for marine environmental protection. In this article, we develop an effective segmentation framework named DGNet, which performs oil spill segmentation by incorporating the intrinsic distribution of backscatter values in SAR images. Specifically, our proposed segmentation network is constructed with two deep neural modules running in an interactive manner, where one is the inference module to achieve latent feature variable inference from SAR images and the other is the generative module to produce oil spill segmentation maps by drawing the latent feature variables as inputs. Thus, to yield accurate segmentation, we take into account the intrinsic distribution of backscatter values in SAR images and embed it in our segmentation model. The intrinsic distribution originates from SAR imagery, describing the physical characteristics of oil spills. In the training process, the formulated intrinsic distribution guides efficient learning of optimal latent feature variable inference for oil spill segmentation. The efficient learning enables the training of our proposed DGNet with a small amount of image data. This is economically beneficial to oil spill segmentation where the availability of oil spill SAR image data is limited in practice. Additionally, benefiting from optimal latent feature variable inference, our proposed DGNet performs accurate oil spill segmentation. We evaluate the segmentation performance of our proposed DGNet with different metrics, and experimental evaluations demonstrate its effective segmentations.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Fast Flood Extent Monitoring With SAR Change Detection Using Google Earth
           Engine

    • Free pre-print version: Loading...

      Authors: Ebrahim Hamidi;Brad G. Peter;David F. Muñoz;Hamed Moftakhari;Hamid Moradkhani;
      Pages: 1 - 19
      Abstract: Flooding is one of the most frequent and disastrous natural hazards triggered by extreme precipitation, high river runoff, hurricane storm surges, and the compounding effects of various flood drivers. This study introduces a new multisource remote sensing approach that leverages both multispectral optical imagery and the weather- and illumination-independent characteristics of synthetic aperture radar (SAR) data to streamline, automate, and map geographically reliable flood inundation extents. Utilizing the near real-time and cloud computing capabilities of Google Earth Engine (GEE), this process facilitates data acquisition and enables large-scale flood monitoring in an expeditious manner. Two major hurricanes along the U.S. Gulf Coast were evaluated: 1) the 2021 Hurricane Ida to the south of New Orleans, LA, USA, and 2) the 2017 Hurricane Harvey to the east of Houston, TX, USA. We devised a change detection and thresholding framework using multitemporal SAR imagery and validated the results with flood extent maps derived from Landsat 8 and Sentinel-2 optical imagery. We demonstrate that constant threshold values for flood extraction from SAR change detection indices are not ubiquitously suitable for all geographies; thus, we outline a heuristic that can be used to select thresholds suitable for specific sites through a fully automated sensitivity analysis. The results indicated high agreement between the SAR and optical imagery (77%–80%), with SAR providing the benefit of under-cloud detection. Furthermore, our results contribute to scaling the SAR approach to produce rapid and accurate information for decision-makers and emergency responders during time-sensitive flood events.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • On the Spatial and Temporal Variations of Primary Production in the South
           China Sea

    • Free pre-print version: Loading...

      Authors: Luping Song;Zhongping Lee;Shaoling Shang;Bangqin Huang;Jinghui Wu;Zelun Wu;Wenfang Lu;Xin Liu;
      Pages: 1 - 14
      Abstract: Primary production (PP) of the South China Sea (SCS) basin area (waters depth deeper than 200 m) is estimated using satellite products, with an overarching goal to reliably characterize the spatial distribution and temporal variation of PP of this important marginal sea. Among the PP models used, the absorption-based model (AbPM) showed better performance ( $R^{2}=0.47$ and $N =39$ ). In comparison, the $R^{2}$ value is 0.26 for a chlorophyll-based model [vertically generalized production model (VGPM)] and 0.15 for the carbon-based model (CbPM). Furthermore, we observed that the PP spatial patterns obtained from these models were similar but disagree on the annual PP magnitude, where VGPM and CbPM, respectively, obtained $sim $ 50% lower and $sim $ 40% higher annual PP compared to that obtained by AbPM. In particular, after analysis using empirical orthogonal functions (EOFs), the upwelling-induced high PP off Luzon (winter) and Vietnam coast (summer) was clearly reflected in the first EOF mode of the AbPM results, and its principal component 1 has shown a decreasing trend for the period of 2003–2019 (−15.0% yr $^{-1}$ for winter, $p < 0.05$ ; −14.7% yr $^{-1}$ for summer, $p < 0.05- ), which reflects the impact of weakening wind and higher sea surface temperature in the SCS. For the results of VGPM and CbPM, however, no strong relationships were found with the main regional oceanographic features. These results suggested that the spatiotemporal variations of SCS PP obtained from AbPM are more reasonable and further highlight the importance of a robust model in reliably capturing large-scale spatiotemporal dynamics of PP in marine environments.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Comprehensive Deep Learning-Based Outlier Removal Method for Multibeam
           Bathymetric Point Cloud

    • Free pre-print version: Loading...

      Authors: Jiawei Long;Hongmei Zhang;Jianhu Zhao;
      Pages: 1 - 22
      Abstract: To address the drawbacks that current multibeam bathymetric outlier removal methods lack repeatability, often require parameter adjustment for different regions, still require a lot of manual labor, and offer limited scalability, a deep learning-based outlier removal method for multibeam bathymetric data is proposed. The method fully considers the multibeam data measurement principles and causes of outliers in multibeam bathymetric data and includes a comprehensive sample augmentation method and an outlier removal model based on a modification of a recently proposed PCPNet architecture. In our extensive evaluation, both on synthetic and real data, our method demonstrates robust outlier removal performance in a variety of marine environments without any parameter adjustment.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Region-Growing-Based Automatic Localized Adaptive Thresholding Algorithm
           for Water Extraction Using Sentinel-2 MSI Imagery

    • Free pre-print version: Loading...

      Authors: Bharath Kumar Reddy Kadapala;Abdul Hakeem K;
      Pages: 1 - 8
      Abstract: Water is a distinct land cover feature on the Earth. Water can easily be extracted from satellite data under known/controlled conditions using simple thresholding techniques. However, these pixel thresholding techniques fail when applied over large regions because they cannot adapt the thresholds based on local variability. Although region-growing methods are available to overcome this issue, there is a limitation in using a global threshold value determined for a satellite scene. Such thresholds fail to adapt to the variations within the scene. Hence, this study proposed a new region-growing-based approach for improved classification of water features using Sentinel-2 multispectral instrument (MSI) data. The significant difference between this approach and other region-growing methods is that this technique uses dynamic thresholding for localized adaption. The local image statistics of the normalized difference water index (NDWI) layer are used for adding new pixels to the region, and the threshold is adjusted for every new pixel added. This technique also minimizes manual intervention in water classification, thus enabling total process automation in near real time. The thresholds are dynamically determined not only for each scene but also for each new pixel being classified for the accurate delineation of the water pixels. A rigorous quality evaluation was done on the resultant water layer. This technique achieved an overall accuracy of 84.04% in challenging conditions.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Temporal Fusion Based 1-D Sequence Semantic Segmentation Model for
           Automatic Precision Side Scan Sonar Bottom Tracking

    • Free pre-print version: Loading...

      Authors: Xiaoming Qin;Ziyin Wu;Xiaowen Luo;Bin Li;Dineng Zhao;Jieqiong Zhou;Mingwei Wang;Hongyang Wan;Xiaolun Chen;
      Pages: 1 - 16
      Abstract: Precision bottom tracking is a core step in the data processing of side scan sonar (SSS), which is of critical value for the quality of the final SSS production. Currently, the automatic precision SSS bottom tracking remains challenging due to the complex noise caused by the measuring environment; especially, the existing methods did not fully exploit the temporal correlations or depended on the hand-crafted setting. Therefore, we proposed a novel temporal fusion-based 1-D sequence semantic segmentation model, TFSSM–1–D, to fuse the temporal correlation features and perform automatic precision SSS bottom tracking. The TFSSM–1–D uses the deep learning (DL) encoder–decoder model for mapping inputs to 1-D semantic label outputs, with the aid of preprocess and temporal fusion modules, to improve the accuracy and robustness. Among them, preprocess module is used to introduce the prior knowledge of bilateral symmetry, which alleviates the defect of long-distance features correlation caused by the inductive bias of convolution neural network (CNN). The temporal fusion consists of the point-wise temporal fusion module (PTFM), and the bi-directional attention propagation module (BAPM) guides the model to explicitly fuse the temporal variation features on different scales. The experimental results demonstrate the effectiveness of TFSSM–1–D, and its mean port offset error (MPOE) reaches 2.7058 on the testing set without downsampling, which is 40% lower than the previous DL based model with single ping input, and other evaluation metrics have also significantly improved. The inference on unseen data shows that TFSSM–1–D can achieve precision bottom tracking with good noise immunity.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Bagged-Tree Machine Learning Model for High and Low Wind Speed Ocean
           Wind Retrieval From CYGNSS Measurements

    • Free pre-print version: Loading...

      Authors: Pin-Hsuan Cheng;Charles Chien-Hung Lin;Y. T. Jade Morton;Shu-Chih Yang;Jann-Yenq Liu;
      Pages: 1 - 10
      Abstract: This article presents two empirical models, the low wind bagged trees (LWBT) and high wind bagged trees (HWBT) ensemble models to estimate ocean surface wind speed using spaceborne Global Navigation Satellite System Reflectometry (GNSS-R) data. The models are empirically trained using NASA’s Cyclone GNSS (CYGNSS) mission level 1 data (version 2.1). The truth label for the LWBT model is the wind speed product derived from European Centre for Medium-Range Weather Forecasts (ECMWF) ERA-5 and Global Data Assimilation System (GDAS), while the label for the HWBT model is wind speed measurements from stepped frequency microwave radiometer (SFMR). Testing results show that the LWBT and HWBT models achieved global wind speed retrieval root-mean-square-error (RMSE) of $sim $ 1.5 and $sim $ 1.4 m/s, respectively, corresponding to an improvement of 29% and 65% with respect to the CYGNSS Level 2 standard wind speed product. The maximum bias is reduced by 65% and 60% for LWBT and HWBT over the Level 2 wind speeds, respectively. Two typhoon case studies are presented to corroborate the model performances and their retrieved wind speeds are consistent with reports from World Meteorological Organization (WMO) and with the measurement provided by the Huangmao Zhou (HMZ) weather station.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • On the Use of Dual Co-Polarized Radar Data to Derive a Sea Surface Doppler
           Model—Part 2: Simulation and Validation

    • Free pre-print version: Loading...

      Authors: Shengren Fan;Biao Zhang;Artem Moiseev;Vladimir Kudryavtsev;Johnny A. Johannessen;Bertrand Chapron;
      Pages: 1 - 9
      Abstract: The Doppler shift obtained from synthetic aperture radar (SAR) measurements comprises the combined contribution to the radial motion of the ocean surface induced by the sea state (wind waves and swell) and underlying surface currents. Hence, to obtain reliable estimates of the ocean surface current (OCS), the sea-state-induced Doppler shifts must be accurately estimated and eliminated. In this study, we use a semiempirical dual co-polarization Doppler velocity (DPDop) model, presented in the companion paper, to calculate sea-state-induced Doppler shifts using buoy-measured wind speed, wind direction, and ocean wave spectra. The DPDop model-simulated Doppler shifts are compared with the collocated Sentinel-1B SAR Wave (WV) mode observations at the 24° and 37° incidence angles, showing a bias of −0.24 Hz and a root-mean-square error (RMSE) of 5.55 Hz. This evaluation is also implemented on a simplified DPDop model at the same incidence angles. The model inputs include wind fields from the European Center for Medium-Range Weather Forecasts (ECMWF) and wave characteristic parameters (e.g., significant wave height (SWH), mean wave direction, and mean wavenumber) from WAVEWATCH III (WW3). The estimated Doppler shifts are validated using the ascending and descending observations of Sentinel-1B WV over the global ocean. Furthermore, the comparisons show that the bias and RMSE are −0.71 and 9.25 Hz, respectively. Based on accurate wave bias correction, we obtain the radial current speeds of the ocean surface from the Doppler shift measurements. The estimated current speeds are compared with the collocated high-frequency (HF) radar measurements, with a bias of −0.04 m/s and an RMSE of 0.15 m/s. These results suggest that the original and simplified DPDop models can be used to estimate sea-state-induced Doppler shifts and, thus, derive accurate surface current retrievals.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Sample Enhancement Method Based on Simple Linear Iterative Clustering
           Superpixel Segmentation Applied to Multibeam Seabed Classification

    • Free pre-print version: Loading...

      Authors: Haiyang Hu;Chengkai Feng;Xiaodong Cui;Kai Zhang;Xianhai Bu;Fanlin Yang;
      Pages: 1 - 15
      Abstract: Full-coverage and high-efficiency seabed sediment detection and identification are critical elements of digital marine construction that support the three-dimensional and thematic development of maritime spatial geographic information systems. With the development of multibeam echo sounder (MBES), the use of MBES backscatter intensity and bathymetry data to extract backscatter angular response (AR) features has increased. Using backscatter intensity features and seabed terrain features for classification is an effective way to achieve large-scale seabed sediment classification. However, it was still limited by small sample size problems and the poor stability of the classification model due to the complexity of performing seabed sediment sampling. In response to the above issues, this article proposes a sample enhancement method based on simple linear iterative clustering (SLIC) superpixel segmentation to address the problems. First, a superpixel-based sample homogeneity expansion method is combined with multibeam backscatter intensity images to achieve adaptive sample range selection. Then, a random forest (RF)-based model is constructed using MBES backscatter intensity and seabed terrain features. To assess the model’s validity, the experiment uses data from an extensive MBES survey and field sampling information from the Celtic Sea, U.K. It achieves accurate predictions for the area’s eight sediment types. The experimental results show that the proposed method expands 60 groups of original sample points to 9293 groups of valid sample points and achieves an overall classification accuracy and kappa coefficient of 84.05% and 0.81 for the seabed sediment, respectively. In addition, the classification accuracy is significantly better compared to the traditional methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multiscale and Multisubgraph-Based Segmentation Method for Ocean Remote
           Sensing Images

    • Free pre-print version: Loading...

      Authors: Qianna Cui;Haiwei Pan;Kejia Zhang;Xiaokun Li;Haoyu Sun;
      Pages: 1 - 20
      Abstract: Interpreting ocean remote sensing images is still a challenge that is worth studying because they can carry valuable information for various important applications. Due to the absence of labeled datasets, unsupervised object-based image analysis (OBIA) methods provide an effective solution to understand remote sensing images with the advantage of grouping local similar pixels into a homogeneous area. However, ocean remote sensing images usually have the characteristics of large size, large background, and coexisting of large and small objects, which results in previous OBIA methods easily falling into the difficulty of accurately segmenting the large and small objects at the same time and the dilemma of time-consuming computation. To solve this problem, a novel multiscale and multisubgraph (MSMSG)-based image segmentation method is presented in this article. First, a coarse-to-fine superpixel generation method is designed to generate optimal superpixels, which can not only solve the problem of coexisting large objects and small objects but also the problem of manually setting the initial segmentation number. Second, the proposed background removal strategy helps to eliminate the trouble of large background areas in ocean remote sensing images. Third, a multisubgraph is constructed with the help of background removal. Finally, the MSMSG merging strategy is addressed to group all similar superpixels into the same cluster, which not only reduces the useless computation of nonadjacent superpixels but also avoids segmentation errors with the same scale. Experiments conducted on three different datasets show that the proposed segmentation method is high-performance and high-efficiency.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Uncertainty Exploration: Toward Explainable SAR Target Detection

    • Free pre-print version: Loading...

      Authors: Zhongling Huang;Ying Liu;Xiwen Yao;Jun Ren;Junwei Han;
      Pages: 1 - 14
      Abstract: Deep learning-based synthetic aperture radar (SAR) target detection has been developed for years, with many advanced methods proposed to achieve higher indicators of accuracy and speed. In spite of this, the current deep detectors cannot express the reliability and interpretation in trusting the predictions, which are crucial especially for those ordinary users without much expertise in understanding SAR images. To achieve explainable SAR target detection, it is necessary to answer the following questions: how much should we trust and why cannot we trust the results. With this purpose, we explore the uncertainty for SAR target detection in this article by quantifying the model uncertainty and explaining the ignorance of the detector. First, the Bayesian deep detectors (BDDs) are constructed for uncertainty quantification, answering how much to trust the classification and localization result. Second, an occlusion-based explanation method (U-RISE) for BDD is proposed to account for the SAR scattering features that cause uncertainty or promote trustworthiness. We introduce the probability-based detection quality (PDQ) and multielement decision space for evaluation besides the traditional metrics. The experimental results show that the proposed BDD outperforms the counterpart frequentist object detector, and the output probabilistic results successfully convey the model uncertainty and contribute to more comprehensive decision-making. Furthermore, the proposed U-RISE generates an attribution map with intuitive explanations to reveal the complex scattering phenomena about which BDD is uncertain. We deem our work will facilitate explainable and trustworthy modeling in the field of SAR image understanding and increase user comprehension of model decisions.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Reducing Biases in Thermal Infrared Surface Radiance Calculations Over
           Global Oceans

    • Free pre-print version: Loading...

      Authors: Nicholas R. Nalli;James A. Jung;Robert O. Knuteson;Jonathan Gero;Cheng Dang;Benjamin T. Johnson;Lihang Zhou;
      Pages: 1 - 18
      Abstract: Thermal infrared (IR) environmental satellite data assimilation and remote sensing of the surface and lower troposphere depend on the accurate specification of the spectral surface emissivity within clear-sky forward calculations. Over ocean surfaces, accurate modeling of surface-leaving radiances over the sensor scanning swaths is complicated by a quasi-specular bidirectional reflectance distribution function (BRDF). Recent findings at the Joint Center for Satellite Data Assimilation (JCSDA) have also revealed significant zonally varying systematic biases ( $approx 0.5 $ K) on a global scale over cold ocean waters; these are the results of temperature dependence in the thermal IR optical constants. This article proposes practical solutions to these problems by modeling thermal IR “effective emissivity” in a manner that accounts for both surface emission and quasi-specular reflectance, along with temperature dependence, while meeting the latency and computational constraints of operational global data assimilation and retrieval systems. We overview the theoretical basis of the model and validate it against ship-based Marine Atmospheric Emitted Radiance Interferometer (MAERI) spectra obtained from cold and warm water ocean campaigns.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • High-Precision Underwater Acoustic Localization of the Black Box Utilizing
           an Autonomous Underwater Vehicle Based on the Improved Artificial
           Potential Field

    • Free pre-print version: Loading...

      Authors: Sibo Sun;Huigong Guo;Guangming Wan;Chao Dong;Ce Zheng;Yong Wang;
      Pages: 1 - 10
      Abstract: Underwater acoustic localization (UWAL) of the black box for a sunken airplane utilizing an autonomous underwater vehicle (AUV) is a useful technique in ensuring traffic safety. Aiming at improving localization precision, this article proposes a new path-planning algorithm based on the improved artificial potential field (APF). Compared with the conventional APF, we modify the conventional gravitation force and introduce a new localization precision force. Therefore, a balance between localization precision and obstacle avoidance is achieved, and the localization precision is significantly improved. The lake trial result validates the effectiveness of the proposed method.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • DensePPMUNet-a: A Robust Deep Learning Network for Segmenting Water Bodies
           From Aerial Images

    • Free pre-print version: Loading...

      Authors: Deqiang Xiang;Xin Zhang;Wei Wu;Hongbin Liu;
      Pages: 1 - 11
      Abstract: The identification of water bodies from aerial images using semantic segmentation networks can provide accurate information for ecological monitoring, flood prevention, and disaster reduction. Outliers on aerial images might reduce interclass separability and thus cause discontinuous prediction of water bodies. The fusion of global context information is helpful to solve this problem. However, the existing global prior representation does not provide sufficient information for identifying a large number of multiscale objects and outliers. In this study, a dense pyramid pooling module (DensePPM) was introduced to extract global prior knowledge with a dense scale distribution. The ablation experiments showed that the models using the DensePPM had higher values of IoU, $F1$ -score, and recall than that using pyramid pooling module (PPM), showing that the proposed module could capture more global context information of outliers under multiscale scenarios. A robust deep learning network named DensePPMUNet-a based on the DensePPM was then proposed for segmenting water bodies from aerial images. The comparative experiments with different datasets demonstrated that the DensePPMUNet-a outperformed other state-of-the-art deep learning models.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Single-Pass Tropical Cyclone Detector and Scene-Classified Wind Speed
           Retrieval Model for Spaceborne GNSS Reflectometry

    • Free pre-print version: Loading...

      Authors: Feng Wang;Guodong Zhang;Dongkai Yang;Hui Kuang;
      Pages: 1 - 16
      Abstract: Spaceborne Global Navigation Satellite System-Reflectometry (GNSS-R) has been used to retrieve wind speed, but few studies have directly detected tropical cyclones from spaceborne GNSS-R observations. Moreover, it is challenging to solve the multivalued dependence of the spaceborne observable on wind speed to unambiguously and accurately retrieve wind speed in cyclone conditions. This article presents a single-pass cyclone detector and coarse estimation approach of the tropical cyclone position using a spaceborne GNSS-R full delay-Doppler map (FDDM). When the specular point passes through a tropical cyclone, the FDDM asymmetry experiences an abnormal change. From the FDDM asymmetry sequence along the specular point trajectory, two subsequence features, including the slope and extremum difference, are defined, from which the tropical cyclone detector is proposed. The results from the simulation and the Cyclone GNSS (CYGNSS) data both indicate a good detection performance for tropical cyclones. The location corresponding to the peak detector is considered a coarse estimation of the tropical cyclone position. The test result from the CYGNSS data shows that the mean error between the detected cyclone center and the International Best Track Archive for Climate Stewardship (IBTrACS) cyclone center is approximately 124.50 km. From the detector of tropical cyclones, a scene-classified wind speed retrieval model is proposed. The simulated and experimental results show that a better retrieval performance can be obtained at high wind speed ( ${>}30$ m/s) using the scene-classified model. Reductions of 10 and 4 m/s in the root mean square errors (RMSEs) are obtained for the simulation and CYGNSS data, respectively. This work is meaningful for directly detecting tropical cyclones and retrieving high wind speed data using spaceborne GNSS-R in real time.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Novel Noise-Aware Deep Learning Model for Underwater Acoustic Denoising

    • Free pre-print version: Loading...

      Authors: Aolong Zhou;Wen Zhang;Xiaoyong Li;Guojun Xu;Bingbing Zhang;Yanxin Ma;Junqiang Song;
      Pages: 1 - 13
      Abstract: Underwater acoustic signal denoising technology aims to overcome the challenge of recovering valuable ship target signals from noisy audios by suppressing underwater background noise. Traditional statistical-based denoising techniques are difficult to be applied effectively in complex underwater environments, especially in the case of extremely low signal-to-noise ratios (SNRs). To address these problems, we propose a noise-aware deep learning model with fullband–subband attention network (NAFSA-Net) for underwater acoustic signal denoising. NAFSA-Net adopts an encoder to extract the feature representation of the input audio. Subsequently, the noise subnet and the target subnet are designed to estimate the noise component and the target component simultaneously. Specifically, some stacked fullband–subband attention (FSA) blocks are deployed in each subnet to capture both global dependencies and fine-grained local dependencies of features. Furthermore, we introduce an interaction module to transmit auxiliary information from the noise subnet to the target subnet. Finally, we propose an improved weight scale-invariant signal-to-noise ratio (SI-SNR) loss function to optimize the training of our model. Experimental results show that our proposed NAFSA-Net substantially outperforms traditional methods and competitive DNN-based solutions in denoising underwater noisy signals with very low SNRs. More importantly, our proposals achieve equally excellent performance on both unseen datasets, which indicates that NAFSA-Net can be a more robust choice for real-world underwater acoustic denoising systems.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Integrating Unordered Time Frames in Neural Networks: Application to the
           Detection of Natural Oil Slicks in Satellite Images

    • Free pre-print version: Loading...

      Authors: Antoine Scardigli;Laurent Risser;Chihab Haddouche;Romain Jatiault;
      Pages: 1 - 14
      Abstract: In this article, we explore the use of novel neural-network architectures to distinguish natural seepages from artificial slicks in synthetic aperture radar (SAR) images. They exploit a distinctive property of natural seepages, which is their temporal recurrence in the same geographical area. This information can be captured in different SAR images acquired at the same location over time, but not necessarily at a regular time frequency. The proposed neural-network architectures are then built as specific block layers, which efficiently treat the unordered temporal information, followed by more conventional neural-network layers, which are widely used for image classification. Different block layers for unordered temporal information are compared on Sentinel-1 images acquired in the Aegean Sea. Following data augmentation steps, our dataset contains a consistent subset of images gathered among 16000 time-patches. We demonstrate that nonlinear time-based block layers and block layers that avoid information bottlenecks are most efficient to discriminate between natural and artificial oil spills. Compared with standard neural-networks which use single time frames, the integration of unordered temporal information increases the overall accuracy (OA) from 82% to 92% on our dataset, which demonstrates the effectiveness of the proposed approach.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Physical Knowledge-Enhanced Deep Neural Network for Sea Surface
           Temperature Prediction

    • Free pre-print version: Loading...

      Authors: Yuxin Meng;Feng Gao;Eric Rigall;Ran Dong;Junyu Dong;Qian Du;
      Pages: 1 - 13
      Abstract: Traditionally, numerical models have been deployed in oceanography studies to simulate ocean dynamics by representing physical equations. However, many factors pertaining to ocean dynamics seem to be ill-defined. We argue that transferring physical knowledge from observed data could further improve the accuracy of numerical models when predicting sea surface temperature (SST). Recently, the advances in Earth observation technologies have yielded a monumental growth of data. Consequently, it is imperative to explore ways to improve and supplement numerical models utilizing the ever-increasing amounts of historical observational data. To this end, we introduce a method for SST prediction that transfers physical knowledge from historical observations to numerical models. Specifically, we use a combination of an encoder and a generative adversarial network (GAN) to capture physical knowledge from the observed data. The numerical model data are then fed into the pretrained model to generate physics-enhanced data, which can then be used for SST prediction. Experimental results demonstrate that the proposed method considerably enhances SST prediction performance compared to several state-of-the-art baselines.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Remote Sensing of Particulate Organic Carbon on the Western Antarctic
           Peninsula Shelf Using a Color Index-Based Algorithm

    • Free pre-print version: Loading...

      Authors: Chengfeng Le;Ming Wu;Sunbin Cai;Na Liu;Jianfeng He;
      Pages: 1 - 13
      Abstract: A color index (CI)–based algorithm (CI $_{mathbf {POC}}$ ) for estimating particulate organic carbon (POC) concentrations from satellite observations is evaluated for the Western Antarctic Peninsula (WAP) shelf. The CI utilizes three spectral bands centered approximately at 490, 550, and 670 nm of the ocean color satellite sensors. The CI $_{mathbf {POC}}$ algorithm is calibrated and validated using field-measured POC concentration data and match-up satellite-derived remote-sensing reflectance ( $text{R}_{mathbf {rs}}$ ) data. For comparison, the performance of the blue–green (BG) band ratio algorithm is also evaluated using the same match-up dataset. Results show that the CI $_{mathbf {POC}}$ algorithm outperforms the BG band ratio algorithm in determining the POC concentration for the WAP shelf with good statistical parameters of algorithm performance for the CI $_{mathbf {POC}}$ algorithm. Given that the CI $_{mathbf {POC}}$ algorithm is built upon a three-band-difference concept and is less sensitive to artificial errors in $R_{mathbf {rs}}$ than the BG algorithm, CI $_{mathbf {POC}}$ can correct POC concentration by retrieving uncertainties caused by sea ice and cloud contaminations. Further evaluation showed that the CI $_{mathbf {POC}}$ algorithm can derive POC from the next-g-neration ocean color satellite sensors with robust performance and noise tolerance.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Toward the Discrimination of Oil Spills in Newly Formed Sea Ice Using
           C-Band Radar Polarimetric Parameters

    • Free pre-print version: Loading...

      Authors: Elvis Asihene;Gary Stern;David G. Barber;Colin Gilmore;Dustin Isleifson;
      Pages: 1 - 15
      Abstract: Climate-driven sea ice loss has exposed the Arctic to increased human activity, which comes along with a higher risk of oil spills. As a result, we investigated the ability of C-band polarimetric parameters in a controlled mesocosm to accurately identify and discriminate between oil-contaminated and uncontaminated newly formed sea ice (NI). Parameters, such as total power, copolarization ratio, copolarization correlation coefficient, and others, were derived from the normalized radar cross section and covariance matrix to characterize the temporal evolution of NI before and after oil spill events. For separation purposes, entropy ( $H$ ) and mean-alpha ( $alpha$ ) were extracted from eigen decomposition of the coherency matrix. The $H$ versus $alpha $ scatterplot revealed that a threshold classifier of 0.3- $H$ and 18°- $alpha $ could distinguish oil-contaminated NI from its oil-free surroundings. From the temporal evolution of the polarimetric parameters, the results demonstrate that the copolarization correlation coefficient is the most reliable polarimetric parameter for oil spill detection, as it provides information on a variety of oil spill scenarios, including oil encapsulated within ice and oil spreading on top of ice. Overall, these findings will be used to support existing and future C-band polarimetric radar satellites for resolving ambiguities associated with Arctic oil spill events, particularly during freeze-up seasons.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Predicting Sea Ice Concentration With Uncertainty Quantification Using
           Passive Microwave and Reanalysis Data: A Case Study in Baffin Bay

    • Free pre-print version: Loading...

      Authors: Xinwei Chen;Ray Valencia;Armina Soleymani;K. Andrea Scott;
      Pages: 1 - 13
      Abstract: In recent years, the adoption of deep learning (DL) techniques for predicting sea ice concentration (SIC) given both passive microwave (PM) data and reanalysis data has seen a growing interest. For use in downstream services, these SIC estimates should be accompanied by uncertainty estimates. To provide these estimates, we utilize a heteroscedastic Bayesian neural network (HBNN), which can estimate both model (epistemic) and data (aleatoric) uncertainty. We use both PM and atmospheric data as our input features and demonstrate that both are needed for accurate SIC estimates. Results show that, over an annual cycle, the months of melt onset, such as April, May, and June, produce the highest uncertainties relative to other months, with total (epistemic + aleatoric) uncertainties of approximately 20%, while areas in the marginal ice zone contributed highest total uncertainty of 25% spatially. When considering an average over the test year, the level of uncertainty due to the data (aleatoric) is consistent with other studies, at 10%–15%. The advantage of our approach is that the uncertainties are specific to the data instance, and both model and data uncertainties are estimated.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Joint Active and Passive Microwave Thermometry of Ice Sheets

    • Free pre-print version: Loading...

      Authors: Anna L. Broome;Dustin M. Schroeder;Joel T. Johnson;
      Pages: 1 - 10
      Abstract: Measurement of ice-sheet thermal state via microwave remote sensing techniques has the potential to provide critically needed observations on englacial temperature at the local and continental scale. Better constraints of the vertical englacial temperature structure are needed to improve understanding of thermomechanical processes, ice rheology, and basal sliding, and to reduce uncertainty in interpretations of basal conditions such as material, roughness, and thermal state. We investigate the potential to combine active and passive microwave remote sensing techniques, namely ice-penetrating radar sounding and microwave radiometry, to enable more precise, accurate, and robust measurement of ice-sheet thermal state across widely varying thermal regimes. We simulate the effects of englacial temperature profiles on the attenuation and brightness temperature and explore the performance tradeoffs for joint radar-radiometer system architectures. Our analysis shows that active and passive microwave measurements have complementary sensitivities to englacial temperature as a function of depth, and that a ground-based joint radar-radiometer system can reduce the requirements and complexity demanded of each single instrument.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A New Sea Ice Concentration Retrieval Algorithm Based on Relationship
           Between AMSR2 89-GHz Polarization and Landsat 8 Observations

    • Free pre-print version: Loading...

      Authors: Yasuhiro Tanaka;Junshen Lu;
      Pages: 1 - 15
      Abstract: A newly developed linear sea ice concentration (SIC) retrieval algorithm based on passive microwave Advanced Microwave Scanning Radiometer 2 (AMSR2) measurements is proposed. SIC is retrieved by a linear function of the polarization ratio (PR) at 89 GHz (PR $_{mathbf {89}}$ ) corrected for atmospheric influence. We use Landsat 8 SIC data to derive the coefficients of the linear function. Results using this linear algorithm are compared to those of ASI2 developed by Lu et al. (2018), which is a nonlinear 89-GHz algorithm with polarization difference (PD) at 89 GHz (PD $_{mathbf {89}}$ ) that also includes a correction for atmospheric influence. Both algorithms are compared with independent SIC data derived from Landsat 8, ship-based observation, and synthetic aperture radar (SAR) and both tend to underestimate the ship-based and Landsat 8 SICs, particularly over thin ice. However, the proposed algorithm tends to provide results with lower bias and root-mean-square error (RMSE) for different ice categories.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Optimization of Cross Correlation Algorithm for Annual Mapping of Alpine
           Glacier Flow Velocities; Application to Sentinel-2

    • Free pre-print version: Loading...

      Authors: Jérémie Mouginot;Antoine Rabatel;Etienne Ducasse;Romain Millan;
      Pages: 1 - 12
      Abstract: Nowadays, satellite observations cover most of the Earth’s surface in a repetitive manner. This information is crucial for documenting variability and environmental changes such as glacier surface velocity. With this in mind, digital image processing has been developed and improved over the past decades. The processing challenges are now related to optimizing parameters that account for the high variability of natural processes, as well as filtering and aggregating the results to provide useful products to end-users. Based on the normalized cross correlation (NCC) method applied to Sentinel-2 optical satellite observations up to 400 days apart, we present a series of tests to derive optimal parameter values for the quantification of alpine glacier ice velocity that we have applied to the Mont-Blanc massif where in situ measurements are available. We found that a search distance adapted to the temporal baseline, a $16times 16$ pixel window size, and a $5times 5$ pixels sampling provide an appropriate combination of parameters to process Sentinel-2 with the NCC method when applied to small alpine glaciers. Combining several spatial and temporal filters applied to a large set of more than 18000 displacement maps obtained between 2015 and 2021, then aggregating these filtered maps using statistical or linear regressions into annual maps, yields near-complete maps of the test region with a root mean square error (RMSE) reduced to about 10 m.yr $^{-1}$ compared to in situ measurements.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Mapping Forest Canopy Height at Large Scales Using ICESat-2 and Landsat:
           An Ecological Zoning Random Forest Approach

    • Free pre-print version: Loading...

      Authors: Zhaocong Wu;Fanglin Shi;
      Pages: 1 - 16
      Abstract: Forest canopy height (FCH) is a crucial indicator in the calculation of forest biomass and carbon sinks. There are various methods to measure FCH, such as space-borne light detection and ranging (LiDAR), but their data are spatially discrete and do not provide continuous FCH maps. Therefore, an FCH estimation method that associates sparse LiDAR data with spatially continuous variables is required. The traditional approach of constructing a single model overlooks the spatial variability in forest growth, which will limit the FCH accuracy. Considering the distinct nature of forest in different ecological zones, the following hold. First, we proposed an ecological zoning random forest (EZRF) model in 33 ecological zones in China. Compared with the total zone RF (TZRF) model, the EZRF model showed a greater potential, which was 21.5%–36.5% more accurate than the TZRF model. Second, we analyzed a total of 62 variables related to forest growth, including Landsat variables and ancillary variables (forest canopy cover, bioclimatic, topographic, and hillshade factors). An insight into variable selection in FCH modeling was provided by analyzing the prediction accuracy of FCH under different categorical variables and analyzing the importance of variables in different ecological zones. Third, finally, we produced a 30-m continuous FCH map by the EZRF model. Compared with the airborne LiDAR data, the FCH prediction results produced a root mean square error (RMSE) of 2.50–5.35 m, which were 41%–72% more precisely than the Global FCH, 2019. The results demonstrate the effectiveness of our proposed method and contribute to the study of forest carbon sinks.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Land Cover Classification With Gaussian Processes Using
           Spatio-Spectro-Temporal Features

    • Free pre-print version: Loading...

      Authors: Valentine Bellet;Mathieu Fauvel;Jordi Inglada;
      Pages: 1 - 21
      Abstract: In this article, we propose an approach based on Gaussian processes (GPs) for large-scale land cover pixel-based classification with Sentinel-2 satellite image time series (SITS). We used a sparse approximation of the posterior combined with variational inference to learn the GP’s parameters. We applied stochastic gradient descent and GPU computing to optimize our GP models on massive datasets. The proposed GP model can be trained with hundreds of thousands of samples, compared to a few thousands for traditional GP methods. Moreover, we included the spatial information by adding the geographic coordinates into the GP’s covariance function to efficiently exploit the spatio-spectro-temporal structure of the SITS. We ran experiments with Sentinel-2 SITS of the full year 2018 over an area of 200000 km2 (about 2 billion pixels) in the south of France, which is representative of an operational setting. Adding the spatial information significantly improved the results in terms of classification accuracy. With spatial information, GP models have an overall accuracy of 79.8. They are more than three points above random forest (the method used for current operational systems) and more than one point above a multilayer perceptron. Compared to a transformer-based model (which provides state-of-the-art results in the literature, but is not applied in operational systems), GP models are only one point below.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Method for Retrieving Coarse-Resolution Leaf Area Index for Mixed Biomes
           Using a Mixed-Pixel Correction Factor

    • Free pre-print version: Loading...

      Authors: Yadong Dong;Jing Li;Ziti Jiao;Qinhuo Liu;Jing Zhao;Baodong Xu;Hu Zhang;Zhaoxing Zhang;Chang Liu;Yuri Knyazikhin;Ranga B. Myneni;
      Pages: 1 - 17
      Abstract: The leaf area index (LAI) is a key structural parameter of vegetation canopies. Accordingly, several moderate-resolution global LAI products have been produced and widely used in the field of remote sensing. However, the accuracy of the current moderate-resolution global LAI products cannot satisfy the requirements recommended by the LAI application communities, especially in heterogeneous areas composed of mixed land cover types. In this study, we propose a mixed-pixel correction (MPC) method to improve the accuracy of LAI retrievals over heterogeneous areas by considering the influence of heterogeneity caused by the mixture of different biome types with the help of high-resolution land cover maps. The DART-simulated LAI, the aggregated Landsat LAI, and the site-based high-resolution LAI reference maps are used to evaluate the performance of the MPC method. The results indicate that the MPC method can reduce the influences of spatial heterogeneity and biome misclassification to obtain the LAI with much better accuracy than the Moderate Resolution Imaging Spectroradiometer (MODIS) main algorithm, given that the high-resolution land cover map is accurate. The root mean square error (RMSE) (bias) decreases from 0.749 (0.486) to 0.414 (0.087), while the R2 increases from 0.084 to 0.524, and the proportion of pixels that fulfill the uncertainty requirement of the GCOS increases from 38.2% to 84.6% for the results of site-based high-resolution LAI reference maps. Spatially explicit information about vegetation fractional cover can further reduce uncertainties induced by variations in canopy density for the results of DART simulated data. The proposed method shows potential for improving global moderate-resolution LAI products.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Passive Only Microwave Soil Moisture Retrieval in Indian Cropping
           Conditions: Model Parameterization and Validation

    • Free pre-print version: Loading...

      Authors: Dileep Kumar Gupta;Prashant K. Srivastava;Dharmendra Kumar Pandey;Sumit Kumar Chaudhary;Rajendra Prasad;Peggy E. O’Neill;
      Pages: 1 - 12
      Abstract: The present study carried out to parameterize the single channel soil moisture active passive (SMAP) passive soil moisture (SM) retrieval algorithm, over Indian conditions. The moderate resolution imaging spectroradiometer (MODIS) data products and soil texture data were used for an improved parameterization of the algorithm. The bias correction was applied to the MODIS leaf area index (LAI) for accurate computation of vegetation optical depth. The necessary vegetation and roughness parameter were calibrated through minimization of the error between model retrieved and ground measured SM. The value of root mean square error (RMSE) for retrieved SM was found as $0.059,,m^{3}m^{-3}$ with bias and correlation coefficients of $0.036,,m^{3}m^{-3}$ and 0.724 for ascending overpass, respectively, while a lower value was recorded (RMSE = $0.059,,m^{3}m^{-3}$ , bias = $0.024,,m^{3}m^{-3}$ , and correlation coefficients = 0.752) for descending overpass. The same method is also implemented on two other test sites in different regions of India to check the model robustness, which indicates that the current parameterization provides a better estimate of SM over croplands in India. The overall performance of new parameterized model is found as (RMSE = 0.052 and bias = 0.034) for ascending and descending (RMSE = 0.048 and bias = 0.026) satellite overpasses for all the three test sites. Additionally, the intercomparing of various operational SM products SMAP SM (L2_SM_P), Soil Moisture and Ocean Salinity (SMOS) SM (SMOS&-x005F;L3_SM), and SMOS-IC data products was carried out with the SAC-ISRO PAN India SM network, which showed a significant RMSE, dry and wet biases over all three test sites as compared to the developed improved parameterized algorithm.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Determination and Evaluation of Surface Solar Irradiance With the
           MAGIC-Heliosat Method Adapted to MTSAT-2/Imager and Himawari-8/AHI Sensors
           

    • Free pre-print version: Loading...

      Authors: Enric Valor;Jesús Puchades;Raquel Niclòs;Joan Miquel Galve Romero;Oriol Lacave;Patrícia Puig;
      Pages: 1 - 19
      Abstract: Surface solar irradiance (SSI) is a crucial component of the radiation budget at the surface, which governs water and energy exchanges with the atmosphere. Good estimates of SSI at regional-to-global scales are needed for modeling land surface processes, climate and weather predictions, or management of solar power plants. This article presents the adaptation of the Mesoscale Atmospheric Global Irradiance Code (MAGIC)-Heliosat method used by the Climate Monitoring Satellite Application Facility (CM-SAF) for Meteosat Second Generation Spinning Enhanced Visible and Infrared Imager (MSG/SEVIRI) to the Multifunction Transport Satellite 2 (MTSAT-2)/Imager and Himawari-8/Advanced Himawari Imager (AHI) sensors managed by the Japanese Meteorological Agency. The method allows providing estimates of global horizontal irradiance (GHI) and direct normal irradiance (DNI) over the Asian Pacific coast and Oceania. These estimates were evaluated by comparison to ground data measured at six baseline surface radiation network (BSRN) stations during years 2014 and 2016. The results showed that GHI can be determined with an accuracy of $-5,,text{W}cdot text{m}^{-2}$ , a precision of 160 $text{W}cdot text{m}^{-2}$ , and a relative absolute error of 30% in an hourly basis. They improved to an accuracy of $-5,,text{W}cdot text{m}^{-2}$ ( $-5,,text{W}cdot text{m}^{-2}$ ), a precision of 70 $text{W}cdot text{m}^{-2}$ (40 $text{W}cdot text{m}^{-2}$ ), and a relative error of 10% (7%) in-daily (monthly) estimates. The results for DNI showed an accuracy of $-45,,text{W}cdot text{m}^{-2}$ and a precision of 330 $text{W}cdot text{m}^{-2}$ , which represent a relative absolute error of 38%. These results improved for longer time steps, with an accuracy of +15 $text{W}cdot text{m}^{-2}$ (+30 $text{W}cdot text{m}^{-2}$ ), a precision of 150 $text{W}cdot text{m}^{-2}$ (130 $text{W}cdot text{m}^{-2}$ ), and a relative error of 35% (20%) in daily (monthly) estimations.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Self-Supervised Global–Local Contrastive Learning for Fine-Grained
           Change Detection in VHR Images

    • Free pre-print version: Loading...

      Authors: Fenlong Jiang;Maoguo Gong;Hanhong Zheng;Tongfei Liu;Mingyang Zhang;Jialu Liu;
      Pages: 1 - 13
      Abstract: Self-supervised contrastive learning (CL) can learn high-quality feature representations that are beneficial to downstream tasks without labeled data. However, most CL methods are for image-level tasks. For the fine-grained change detection (FCD) tasks, such as change or change trend detection of some specific ground objects, it is usually necessary to perform pixel-level discriminative analysis. Therefore, feature representations learned by image-level CL may have limited effects on FCD. To address this problem, we propose a self-supervised global–local contrastive learning (GLCL) framework, which extends the instance discrimination task to the pixel level. GLCL follows the current mainstream CL paradigm and consists of four parts, including data augmentation to generate different views of the input, an encoder network for feature extraction, a global CL head, and a local CL head to perform image-level and pixel-level instance discrimination tasks, respectively. Through GLCL, features belonging to different perspectives of the same instance will be pulled closer, while features of different instances will be alienated, which can enhance the discriminativeness of feature representations from both global and local perspectives, thereby facilitating downstream FCD tasks. In addition, GLCL makes a targeted structural adaptation to FCD, i.e., the encoder network is undertaken by the common backbone networks of FCD, which can accelerate the deployment on downstream FCD tasks. Experimental results on several real datasets show that compared with other parameter initialization methods, the FCD models pretrained by GLCL can obtain better detection performance.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Curvelet Adversarial Augmented Neural Network for SAR Image Classification

    • Free pre-print version: Loading...

      Authors: Yake Zhang;Fang Liu;Licheng Jiao;Shuyuan Yang;Lingling Li;Meijuan Yang;Jianlong Wang;Xu Liu;
      Pages: 1 - 17
      Abstract: Convolutional neural networks (CNNs) have superior feature learning capabilities with large numbers of labeled samples. The reality is that labeling these samples is costly in terms of human labor. Existing data augmentation methods alleviate the scarcity of labeled samples. However, these methods are not suitable for synthetic aperture radar (SAR) images, owing to special imaging mechanisms and observational objects. The generative SAR images by existing augmented methods show structure distortion. To address this issue, we introduce a curvelet adversarial augmented neural network (CA2NN) for SAR image classification. Specifically, an $text{A}^{2}$ NN is established, which consists of two generative streams and one discriminative stream. In the generative stream, through the mutual transformation between the whole and partial images, more new samples with structural consistency are generated to augment the limited labeled data. In the discriminative stream, these generated samples show certain appearance variations after adversarial training based on the novel joint discriminant criterion. Simultaneously, given the multiscale and multidirectional nature of SAR images, we construct discretized curvelet in 2-D space, aiming to extract the singularity features and avoid overfitting. By integrating curvelet kernels into $text{A}^{2}$ NN, CA2NN can automatically generate more representative features adapting to complex terrain, while greatly reducing the complexity of the network. Experiments are conducted on the SAR images with large-scale and complex scenes, suggesting that the proposed approach significantly improves the classification performance with few labeled samples.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Spatial Hierarchical Reasoning Network for Remote Sensing Visual
           Question Answering

    • Free pre-print version: Loading...

      Authors: Zixiao Zhang;Licheng Jiao;Lingling Li;Xu Liu;Puhua Chen;Fang Liu;Yuxuan Li;Zhicheng Guo;
      Pages: 1 - 15
      Abstract: For visual question answering on remote sensing (RSVQA), current methods scarcely consider geospatial objects typically with large-scale differences and positional sensitive properties. Besides, modeling and reasoning the relationships between entities have rarely been explored, which leads to one-sided and inaccurate answer predictions. In this article, a novel method called spatial hierarchical reasoning network (SHRNet) is proposed, which endows a remote sensing (RS) visual question answering (VQA) system with enhanced visual–spatial reasoning capability. Specifically, a hash-based spatial multiscale visual representation module is first designed to encode multiscale visual features embedded with spatial positional information. Then, spatial hierarchical reasoning is conducted to learn the high-order inner group object relations across multiple scales under the guidance of linguistic cues. Finally, a visual-question (VQ) interaction module is employed to learn an effective image–text joint embedding for the final answer predicting. Experimental results on three public RS VQA datasets confirm the effectiveness and superiority of our model SHRNet.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Domain Adaptation Method for Land Use Classification Based on Improved
           HR-Net

    • Free pre-print version: Loading...

      Authors: Zezhong Zheng;Shuang Yu;Shaobin Jiang;
      Pages: 1 - 11
      Abstract: In recent years, the recognition accuracy of a semantic segmentation model on natural images can yield a very high level. Thus, it is of great significance to utilize semantic segmentation algorithm to obtain land use classification with remote sensing images. However, due to the large differences between natural images and remote sensing images, the standard semantic segmentation algorithm is not effective for land use classification of remote sensing images. In this article, the structure of high-resolution network (HR-Net) algorithm is improved according to the difference between the two kinds of images to make it more suitable for remote sensing images. Furthermore, in order to overcome the dependence of the semantic segmentation algorithm on a large number of high-quality prior data sets, some research experiments are conducted with the improved HR-Net domain adaptation model, and both of the adversarial domain adaptation model and the fusion domain adaptation model based on improved HR-Net and CycleGAN are designed to reduce the workload of manually labeling data. The extensive experimental results show that the classification of our improved HR-Net algorithm and the two domain adaptation models outperform other algorithms that demonstrates the effectiveness and superiority of our algorithms.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Change Detection on Remote Sensing Images Using Dual-Branch Multilevel
           Intertemporal Network

    • Free pre-print version: Loading...

      Authors: Yuchao Feng;Jiawei Jiang;Honghui Xu;Jianwei Zheng;
      Pages: 1 - 15
      Abstract: Change detection (CD) of remote sensing (RS) images is mushrooming up accompanied by the on-going innovation of convolutional neural networks (CNNs). Yet with the high-speed technology upgrade, the obstacle that identifies unbalanced variations in foreground–background categories still lies on the table, especially in cases with limited samples and massive interference such as seasonal turnover, illumination intensity, and building reformation. Moreover, to date, neither of the off-the-shelf methods probes the feasibility of direct interaction between bitemporal images before accessing difference features. In this article, we propose a dual-branch multilevel intertemporal network (DMINet) to efficiently and effectively derive the change representations. Specifically, by unifying self-attention (SelfAtt) and cross-attention (CrossAtt) in a single module, we present an intertemporal joint-attention (JointAtt) block to steer the global feature distribution of each input, motivating information coupling between intralevel representations and meanwhile suppressing the task-irrelevant interferences. In addition, centering more on the detection of difference features, a reliable architecture is designed by spotlighting two concerns, i.e., the difference acquisition using subtraction and concatenation as well as the multilevel difference aggregation using incremental feature alignment. Based on a naive backbone without sophisticated structures, i.e., ResNet18, our model outperforms other state-of-the-art (SOTA) methods on four CD datasets, especially in cases with rarely samples. Moreover, the achievement is attained with light overheads.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • When Multigranularity Meets Spatial–Spectral Attention: A Hybrid
           Transformer for Hyperspectral Image Classification

    • Free pre-print version: Loading...

      Authors: Er Ouyang;Bin Li;Wenjing Hu;Guoyun Zhang;Lin Zhao;Jianhui Wu;
      Pages: 1 - 18
      Abstract: The transformer framework has shown great potential in the field of hyperspectral image (HSI) classification due to its superior global modeling capabilities compared to convolutional neural networks (CNNs). To utilize the transformer to model spatial–spectral information, a hybrid transformer that integrates multigranularity tokens and spatial–spectral attention (SSA) is proposed. Specifically, a token generator is designed to embed the multigranularity semantic tokens, which contributes richer image features to the model by exploiting CNN’s local representation capability. Moreover, a transformer encoder with an SSA mechanism is proposed to capture the global dependencies between different tokens, enabling the model to focus on more differentiated channels and spatial locations to improve the classification accuracy. Ultimately, adaptive weighted fusion is applied to different granularity transformer branches to boost HybridFormer’s classification performance. Experiments were conducted on four new challenging datasets, and the results indicate that HybridFormer achieves state-of-the-art results in terms of classification performance. The code of this work will be available at https://github.com/zhaolin6/HybridFormer for the sake of reproducibility.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Deep Feature Enhancement Method for Land Cover With Irregular and Sparse
           Spatial Distribution Features: A Case Study on Open-Pit Mining

    • Free pre-print version: Loading...

      Authors: Gaodian Zhou;Jiahui Xu;Weitao Chen;Xianju Li;Jun Li;Lizhe Wang;
      Pages: 1 - 20
      Abstract: Land cover classification in mining areas (LCMA) is essential for the environmental assessment of mines and plays a crucial role in their sustainable development. The shapes of mine land occupation elements are irregular, and the overall proportion of their area is relatively small. Therefore, their features may be easily lost during feature extraction, which limits the interpretation accuracy in mining areas. This study attempts to address these issues. We propose a model named EG-UNet to enhance the features of elements with few samples and to capture long-range information. The proposed EG-UNet includes two main modules. First, the edge feature enhancement module, the edges of elements of mine land occupation contain more information than other spatial locations. Hence, during the feature extraction of elements, a Sobel operator is used to extract the object boundary, which increases the weight of these features before the pooling operation for their preservation. Second, the long-range information extraction module, long-range information helps extract tiny objects, such as dumping grounds in the mining area. We present a graph convolutional network (GCN) to capture the long-range features and apply convolutional neural networks to learn the graph construction. A total of ten deep-learning networks were compared using the LCMA semantic segmentation dataset. Our model exhibited the best performance, especially in classifying classes with few samples. Furthermore, to evaluate the general ability of EG-UNet, a benchmark-Gaofen Image Dataset (GID) was used, and the result still reflected the superiority of our method.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multisource Maximum Predictor Discrepancy for Unsupervised Domain
           Adaptation on Corn Yield Prediction

    • Free pre-print version: Loading...

      Authors: Yuchi Ma;Zhengwei Yang;Zhou Zhang;
      Pages: 1 - 15
      Abstract: Recently, with the advent of satellite missions and artificial intelligence techniques, supervised machine learning (ML) methods have been more and more used for analyzing remote sensing (RS) observation data for crop yield prediction. However, due to the domain shift between heterogeneous regions, supervised ML models tend to have poor spatial transferability. As a result, models trained with labeled data from one spatial region (i.e., source domain) often lose their validity when directly applied to another region (i.e., target domain). To address this issue, we proposed a multisource maximum predictor discrepancy (MMPD) neural network that is an unsupervised domain adaptation (UDA) approach for corn yield prediction at the county level. The novelties of this study include that: 1) we proposed to maximize the discrepancy between two source-specific yield predictors and align source and target domains by considering crop yield response in the target domain and 2) we adopted the strategy of multisource UDA to avoid negative interference between labeled samples from different sources. Case studies in the U.S. corn belt and Argentina demonstrated that the proposed MMPD model had effectively reduced domain shifts and outperformed several other state-of-the-art deep learning (DL) and UDA methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Two-Stream Translating LSTM Network for Mangroves Mapping Using Sentinel-2
           Multivariate Time Series

    • Free pre-print version: Loading...

      Authors: Zhaohui Xue;Siyu Qian;
      Pages: 1 - 16
      Abstract: Monitoring mangroves is critical to protect the coastal ecosystems, and deep learning has gained great popularity in mapping mangroves using remote sensing. However, mangroves are usually submerged by cyclical tide since they are grown in land–sea interface places, resulting in some drawbacks of existing mangroves mapping models. On one side, the correlations between the vegetation index (VI) and the water index (WI) time series of mangroves are not fully considered. On another side, existing models rarely explored the local differences between mangroves and other land covers. Considering the above two aspects, we propose a novel two-stream translating long short-term memory network (TSTLN) for mangroves mapping. First, we construct multivariate time series (MTS) by compositing VI and WI based on Sentinel-2 time-series data. Second, we build a two-stream architecture and design a Siamese translating (ST) module in both streams. In the global stream, MTS is embedded into the ST module directly to get global features, whereas, in the local stream, a depthwise convolutional self-attention (DCA) module is conceived to capture local information first, and then, local features are further learned by the ST module. Finally, a fully connected layer and softmax are used to classify the representations extracted from the two streams. Experiments conducted over the Maowei Sea, the Dongzhai Port, and the Quanzhou Bay in 2019 demonstrate that: 1) TSTLN outperforms other methods, with improved OA of 0.49%–3.89%, 1.35%–6.85%, and 0.73%–3.65% in the three areas, respectively; 2) two-stream architecture, ST module, and DCA module all contribute to the good performance of TSTLN; and 3) TSTLN maintains higher accuracy with few parameters and less running time compared to ot-er counterparts.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Normalized Spectral Angle Index for Estimating the Probability of
           Viewing Sunlit Leaves From Satellite Data

    • Free pre-print version: Loading...

      Authors: Meihong Fang;Weimin Ju;Jing M. Chen;Weiliang Fan;Wei He;Feng Qiu;Xiangyan Hu;Jing Li;
      Pages: 1 - 19
      Abstract: The probability of viewing sunlit leaves (PT) is a crucial variable influencing observed canopy spectra. Proper determination of PT is necessary for the quantitative retrieval of vegetation parameters using remote sensing. This article describes a spectral index for estimating PT from satellite-observed canopy spectra. For this purpose, we propose a normalized spectral angle index (NSAI) at near-infrared (NIR) wavelengths, based on the spectral shapes of leaf and soil background. The performance of NSAI in estimating PT was evaluated using one ground-based high-resolution imaging dataset, one synthetic satellite dataset, and one satellite-ground synchronous observation dataset. The results demonstrate that NSAI is more suitable for estimating PT from satellite data than five commonly used spectral indices, including enhanced vegetation index (EVI), normalized difference spectral index (NDSI), normalized difference vegetation index (NDVI), simple ratio (SR) index, and photochemical reflectance index (PRI). NSAI exhibits a significant linear correlation with PT. The empirical model for estimating PT based on NSAI has the best transferability from simulated to in situ satellite data. For the fine spectral–spatial resolution (Hyperion) data, the normalized root-mean-square error (nRMSE) and adjusted $R^{2}$ of estimated PT were 14.9% and 0.744, respectively. For MODIS images, PT was estimated with satisfactory accuracy, with an nRMSE of 18.71% and an adjusted $R^{2}$ of 0.670. NSAI is potentially applicable to satellite images for direct estimation of PT to improve the inversion accuracy of vegetation parameters.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Exploring the Potential of Gaofen-1/6 for Crop Monitoring: Generating
           Daily Decametric-Resolution Leaf Area Index Time Series

    • Free pre-print version: Loading...

      Authors: Baodong Xu;Haodong Wei;Zhiwen Cai;Jingya Yang;Zhewei Zhang;Cong Wang;Jing Li;Jing Zhao;Yonghua Qu;Gaofei Yin;Aleixandre Verger;
      Pages: 1 - 14
      Abstract: High spatiotemporal resolution time series of leaf area index (LAI) are essential for monitoring crop dynamics and validating coarse-resolution LAI products. The optical satellite sensors at decametric resolution have historically suffered from a long revisit cycle and cloud contamination issues that hampered the acquisition of frequent and high-quality observations. The 16-m/four-day resolution of the new-generation Gaofen-1 (GF-1) and Gaofen-6 (GF-6) satellites provide an unprecedented opportunity to address these limitations. Here, we developed an effective strategy to generate daily 16-m LAI maps combining GF-1/6 data and ground LAINet measurements. All high-quality GF-1/6 observations were utilized first to derive smoothed time series of vegetation indices (VIs). Second, a random forest regression (RF-r) model was trained to link the VIs with corresponding field LAI measurements. The trained RF-r was finally employed to generate the LAI maps. Results demonstrated the reliability of the reconstructed daily VIs (relative error (RE) < 1%) and the derived LAI time series, which greatly benefited from GF-1/6 high-frequency observations. The direct comparison with field LAI measurements by LAI-2200/LI-3000 showed the good performance of retrieved LAI maps, with bias, root mean square error (RMSE), and $R^{mathbf {2}}$ of 0.05, 0.59, and 0.75, respectively. The LAI time series well captured the spatiotemporal variation of crop growth. Furthermore, the continuous GF-1/6 LAI maps outperformed Sentinel-2 LAI estimates both in terms of temporal frequency and accuracy. Our study indicates the potential of GF-1/6 to generate continuous decametric-resolution LAI maps for fine-scale agricultural monitoring.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Novel Method for Hidden Natural Caves Characterization and Accessibility
           Assessment From Spaceborne VHR SAR Images

    • Free pre-print version: Loading...

      Authors: Leonardo Carrer;Davide Castelletti;Riccardo Pozzobon;Francesco Sauro;Lorenzo Bruzzone;
      Pages: 1 - 11
      Abstract: Caves are one of the last frontiers of human exploration on Earth. They are very relevant scientific targets as they host significant biodiversity and unique geologic formations. The presence of underground passages accessible for human or robotic exploration are revealed by localized collapse of the near-surface ceiling of a cave system (skylight). Remote sensing systems are a valuable tool for skylights detection as these features are often located on very remote and often inaccessible regions of the Earth. However, with the available remote sensing techniques and data analysis methodologies, it is very difficult to determine whether a skylight is providing access to a cave continuation or it represents only a closed depression with no extensions. In this article we propose a methodology, based on very high-resolution (VHR) orbital synthetic aperture radar (SAR) imaging systems, to estimate both caves geometric characteristics and accessibility information in the proximity of a skylight. To test our methodology, we acquired radar data over different Earth’s location by exploiting the Capella Space X-band microsatellite radar constellation. The experimental results show that our methodology effectively determines the caves geometric characteristics and accessibility under a variety of surface conditions. We also detected several unknown and unexplored large cave systems located near Volcan Wolf and Ecuador, Isla Isabela, Galapagos. The presented work has relevant implications for the field of geological studies, ecology, and space exploration research since optical imaging shows the evidence of potential cave systems accessible from skylights on other planetary bodies such as Mars.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • An Efficient EM Modeling Scheme for Large 3-D Models—A
           Magnetotelluric Case Study

    • Free pre-print version: Loading...

      Authors: Arun Singh;Rahul Dehiya;
      Pages: 1 - 11
      Abstract: We present an efficient scheme for computing 3-D magnetotelluric (MT) forward responses. The scheme is especially valuable for large models resulting from fine discretization or the large survey area. The proposed approach overcomes the iterative solvers’ slow convergence that occurs in large modeling problems due to a sizeable ill-conditioned system matrix that needs to be solved. Primarily, the slow convergence arises due to the grid stretching that is necessary to apply the boundary conditions (BCs). Our approach partly removes the grid stretching, thus improving the computational efficiency. In this scheme, a model is represented using two different meshes. One is a coarse mesh with grid stretching, and another is a fine mesh of the desired discretization excluding grid stretching. Using the electric field computed for the coarse mesh, a radiation boundary (RB) vector is calculated at the outer boundary of the fine mesh and is used to compute the necessary BCs along with an initial guess to be utilized by the iterative solver for the fine mesh. The RB vector can be computed at any arbitrarily shaped interface, thus allowing more flexibility in the shape of the fine mesh boundary. It is a significant advantage compared to the traditional finite difference (FD)-based algorithms where the boundaries must be same as the cuboid surfaces. Through different resistivity models, both synthetic and real, we demonstrate that the proposed approach improves the computational efficiency without compromising the accuracy of the solution while providing more flexibility in the shape of the fine mesh.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Feature-Fusion Segmentation Network for Landslide Detection Using
           High-Resolution Remote Sensing Images and Digital Elevation Model Data

    • Free pre-print version: Loading...

      Authors: Xinran Liu;Yuexing Peng;Zili Lu;Wei Li;Junchuan Yu;Daqing Ge;Wei Xiang;
      Pages: 1 - 14
      Abstract: Landslide is one of the most dangerous and frequently occurred natural disasters. The semantic segmentation technique is efficient for wide area landslide identification from high-resolution remote sensing images (HRSIs). However, considerable challenges exist because the effects of sediments, vegetation, and human activities over long periods of time make visually blurred old landslides very challenging to detect based upon HRSIs. Moreover, for terrain features like slopes, aspect and altitude variations cannot be sufficiently extracted from 2-D HRSIs but can be from digital elevation model (DEM) data. Then, a feature-fusion based semantic segmentation network (FFS-Net) is proposed, which can extract texture and shape features from 2-D HRSIs and terrain features from DEM data before fusing these two distinct types of features in a higher feature layer. To segment landslides from background, a multiscale channel attention module is purposely designed to balance the low-level fine information and high-level semantic features. In the decoder, transposed convolution layer replaces original mathematical bilinear interpolation to better restore image resolution via learnable convolutional kernels, and both dropout and batch normalization (BN) are introduced to prevent over-fitting and accelerate the network convergence. Experimental results are presented to validate that the proposed FFS-Net can greatly improve the segmentation accuracy of visually blurred old landslides. Compared to U-Net and DeepLabV3+, FFS-Net can improve the mean intersection over union (mIoU) metric from 0.508 and 0.624 to 0.67, the F1 metric from 0.254 and 0.516 to 0.596, and the pixel accuracy (PA) metric from 0.874 and 0.906 to 0.92, respectively. For the detection of visually distinct landslides, FFS-NET also offers comparable detection performance, and the segmentation is improved for visually distinct landslides with similar color and texture to surroundings.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multiform Ensemble Self-Supervised Learning for Few-Shot Remote Sensing
           Scene Classification

    • Free pre-print version: Loading...

      Authors: Jianzhao Li;Maoguo Gong;Huilin Liu;Yourun Zhang;Mingyang Zhang;Yue Wu;
      Pages: 1 - 16
      Abstract: Self-supervised learning is an effective way to solve model collapse for few-shot remote sensing scene classification (FSRSSC). However, most self-supervised contrastive learning auxiliary tasks perform poorly on the high interclass similarity problem in FSRSSC. Furthermore, it is time-consuming and computationally expensive to obtain the best combination among numerous self-supervised auxiliary tasks. In practical applications, we may encounter difficulties in remote sensing data acquisition and labeling, while most FSRSSC studies only focus on the former. To alleviate the above problems, we propose a multiform ensemble self-supervised learning (MES2L) framework for FSRSSC in this article. Based on the transfer learning-based few-shot scheme, we design a novel global–local contrastive learning auxiliary task to solve the low interclass separability problem. The self-attention mechanism is designed in the local contrast features to investigate the intrinsic associations between different remote sensing scene objectives. We also present a multiform ensemble enhancement (MEE) training method. Ensemble enhancement involves the concatenation of features extracted from different backbones trained by a combination of multiform self-supervised auxiliary tasks. MEE can not only be regarded as a more straightforward alternative to knowledge distillation but also can achieve an effective compromise between expensive computational cost and classification accuracy. In addition, we provide two scene classification schemes of inductive and transductive settings, corresponding to solving the difficulties of remote sensing data acquisition and labeling. The proposed network achieves state-of-the-art results on three benchmark FSRSSC datasets. The potential of the MES2L framework is also demonstrated in combination with classical metalearning-based and metric learning-based few-shot algorithms.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Seismic Inversion Based on Acoustic Wave Equations Using Physics-Informed
           Neural Network

    • Free pre-print version: Loading...

      Authors: Yijie Zhang;Xueyu Zhu;Jinghuai Gao;
      Pages: 1 - 11
      Abstract: Seismic inversion is a significant tool for exploring the structure and characteristics of the underground. However, the conventional inversion strategy strongly depends on the initial model. In this work, we employ the physics-informed neural network (PINN) to estimate the velocity and density fields based on acoustic wave equations. In contrast to the traditional purely data-driven machine learning approaches, PINNs leverage both available data and the physical laws that govern the observed data during the training stage. In this work, the first-order acoustic wave equations are embedded in the loss function as a regularization term for training the neural networks. In addition to the limited amount of measurements about the state variables available at the surface being used as the observational data, the well logging data is also used as the direct observational data about the model parameters. The numerical results from several benchmark problems demonstrate that given noise-free or noisy data, the proposed inversion strategy is not only capable of predicting the seismograms, but also estimating the velocity and density fields accurately. Finally, we remark that although the absorbing boundary conditions are not imposed in the proposed method, the reflected waves do not appear from the artificial boundary in the predicted seismograms.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Seismic Data Enhancement Based on Common-Reflection-Surface-Based Local
           Slope and Trimmed Mean Filter

    • Free pre-print version: Loading...

      Authors: Chuangjian Li;Suping Peng;Xiaoqin Cui;Wenfeng Du;Peng Lin;
      Pages: 1 - 9
      Abstract: Seismic data often contain noise that can disturb or mask effective information. Noise elimination is an important and challenging task in seismic signal processing. Considering the high amplitude continuity of seismic events in the shot domain, this article proposes a structure-oriented denoising method that can enhance the effective events and suppress disturbing noise, including both incoherent and coherent noise. Based on the common-reflection-surface (CRS) travel time, the local slope of seismic events in the shot domain is deduced and estimated to provide structural information for plane-wave prediction. The proposed CRS-based slope depends on fewer parameters (two in 2-D) than the conventional full CRS travel time (three in 2-D), making it computationally efficient. Using the local slope, the third dimension is created using the plane-wave differential equation to predict the current trace from its neighbor traces and trimmed mean filtering (TMF) is applied in this dimension. The added dimension can be regarded as flattening the seismic events within a neighboring window and collapsing after the application of TMF. Synthetic and field datasets are employed to demonstrate the effectiveness of the proposed structure-oriented TMF. Compared with the wavelet and plane-wave destruction (PWD) methods, the proposed method can preserve more useful information with greater continuity in amplitude.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • An Intelligent MT Data Inversion Method With Seismic Attribute Enhancement

    • Free pre-print version: Loading...

      Authors: Hongyu Zhou;Rui Guo;Maokun Li;Fan Yang;Shenheng Xu;Maoshan Chen;Yongtao Wang;Deqiang Tao;Zuzhi Hu;Xianwen Cui;Qinian Wang;Jiangbo Zhu;Suhe Huang;
      Pages: 1 - 14
      Abstract: Magnetotelluric (MT) data inversion reconstructs an electrical resistivity structure most compatible with the observed MT data, and static correction can remove the undesired static shift effect in MT data. Conventional MT data static shift correction often faces the challenge of demanding requirements, such as large data amount, additional types of data, or a deep understanding of the research area. MT inversion constrained by seismic data often has better resolution and model consistency compared with independent MT inversion. However, valuable inversion knowledge contained in geophysicists’ expertise is not effectively incorporated. In this work, we present an intelligent MT data inversion method leveraging data- and physics-driven techniques based on deep learning. A novel MT data static shift correction method is introduced based on a neural network (NN). An MT data inversion method is formulated with the constraint of the extracted seismic reflection image based on two different NNs. Experiments on synthetic and field data verify the effectiveness of the proposed method.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Modeling the Effect of Multiscale Heterogeneities on Wave Attenuation and
           Velocity Dispersion

    • Free pre-print version: Loading...

      Authors: Yurong Wang;Zhaoyun Zong;Qianhao Sun;
      Pages: 1 - 17
      Abstract: Wave attenuation and velocity dispersion are of great significance to fluid evaluation in fluid-saturated rocks. Inspired by the double-porosity and cracked porous elastic wave theory, we develop a multiscale heterogeneities elastic wave theory to better describe the attenuation characteristics caused by wave-induced fluid flow at various scales in saturated rocks. First, we introduce penny-shaped and narrow-shaped coupled cracks into double-porosity model. The potential energy function is given by the stress–strain relationship. The kinetic energy function and dissipation function are derived based on the generalized Biot’s theory. Next, we derive the multiscale wave equations through the Lagrange equation and obtain three compressional waves and one shear wave. According to the novel wave equations, we work out the phase velocity and inverse quality factor in double-porosity and cracked media based on plane-wave analysis. The numerical simulation results show that there are four attenuation peaks in the whole frequency band, corresponding to mesoscopic fluid flow, and two kinds of squirt flow and Biot flow, respectively. Then, we investigate the effect of poroelastic parameters on wave propagation. It is found that porosity, inclusion size, crack aspect ratio, and other parameters significantly affect the dispersion and attenuation. Finally, we compare experimental data and theoretical prediction. Moreover, the new elastic wave theory can degenerate into classical Biot’s theory, double-porosity theory, and pore-crack theory under certain conditions, which demonstrates the consistency of this method. The proposed theory can better predict and simulate the wave propagation by incorporating the effects of microscopic, mesoscopic, and macroscopic heterogeneities.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Imaging Steeply Dipping Faults Using Angle-Controlled Decoupled Elastic
           Reverse-Time Migration of Multicomponent Seismic Data

    • Free pre-print version: Loading...

      Authors: Hao Hu;Yingcai Zheng;Lianjie Huang;Kai Gao;
      Pages: 1 - 8
      Abstract: High-angle and steeply dipping faults are important for geothermal fluid flow and production. However, imaging such faults is challenging for conventional seismic imaging methods using acoustic waves because of the limited illumination of compressional waves. We develop a novel imaging method to use elastic waves including both compressional-to-compressional (P–P) waves and compressional-to-shear (P-to-S) converted waves to image the high-angle faults. P–S converted waves provide additional seismic illumination for high-angle faults because they have smaller reflection angles than those of the P–P reflection waves at interfaces. Our new imaging method employs the decoupled wave equations in the context of elastic reverse-time migration (ERTM) to perform imaging using both P–P and P–S waves. Our decoupled ERTM does not require the polarity correction of S waves. In addition, we use the Poynting vector, measuring the wave propagation direction by computing the wave energy flux to control the image dip angles, thus improving the ability to image high-angle faults. Based on the angle-controlled P–P and P–S images, we further enhance the structural images of high-angle faults by extracting the significant structures that coexist in both P–P and P–S images. We apply our angle-controlled decoupled ERTM angle-controlled decoupled elastic reverse-time migration (ADEM) to synthetic multicomponent seismic data for a simple layer model with two vertical faults and a modified Marmousi2 model with artificially embedded vertical faults. Our numerical results demonstrate that our ADEM method can effectively image high-angle faults in complex structures using both P–P and P–S waves.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Weakly Supervised Transfer Learning Approach for Radar Sounder Data
           Segmentation

    • Free pre-print version: Loading...

      Authors: Miguel Hoyo García;Elena Donini;Francesca Bovolo;
      Pages: 1 - 18
      Abstract: Airborne radar sounders (RSs) are active sensors that acquire subsurface data for Earth observation. RS data (radargrams) provide information on buried geology by imaging subsurface dielectric discontinuities. Recently, several automatic RS target identification techniques have been proposed, with convolutional neural network (CNN)-based methods being the most promising. However, they require numerous labeled data that are hard to retrieve in the subsurface environment targeted by RS. Furthermore, they are not designed to effectively deal with problems showing unbalanced classes, such as RS segmentation. We introduce newer cryosphere subsurface targets in the inland and coastal areas that can have a very low probability. To deal with the higher complexity and variability than previous works, we propose a transfer learning framework for RS data to mitigate the need for a large amount of labeled data and handle extremely unbalanced target classes. Herewith, we propose two transfer learning-based mechanisms for radargram segmentation. The first uses a lightweight architecture whose pretraining is supervised with a large labeled dataset from other domains. The second mechanism uses a deep architecture pretrained in the RS domain, considering the pretest task of radargram reconstruction. The architectures are modified to deal with the characteristics of RS data and the radargram segmentation task. Finally, both methods are fine-tuned with a few labeled radargrams to learn radargram features useful for segmentation. We reveal experimental results on radargrams acquired in Antarctica by MCoRDS-1 and MCoRDS-3. The results demonstrate the effectiveness of transfer learning for radargram segmentation.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • RockFormer: A U-Shaped Transformer Network for Martian Rock Segmentation

    • Free pre-print version: Loading...

      Authors: Haiqiang Liu;Meibao Yao;Xueming Xiao;Yonggang Xiong;
      Pages: 1 - 16
      Abstract: Martian rock segmentation aims to separate rock pixels from background, which plays a crucial role in downstream tasks, such as traversing and geologic analysis by Mars rovers. The U-Nets have achieved certain results in rock segmentation. However, due to the inherent locality of convolution operations, U-Nets are inadequate in modeling global context and long-range spatial dependencies. Although emerging Transformers can solve this, they suffer from difficulties in extracting and retaining sufficient low-level local information. These shortcomings limit the performance of the existing networks for Martian rocks that are variable in shape, size, texture, and color. Therefore, we propose RockFormer, the first U-shaped Transformer framework for Mars rock segmentation, consisting of a hierarchical encoder–decoder architecture with a feature refining module (FRM) connected between them. Specifically, the encoder hierarchically generates multiscale features using an improved vision Transformer (improved-ViT), where both abundant local information and long-range contexts are exploited. The FRM removes less representative features and captures global dependencies between multiscale features, improving RockFormer’s robustness to Martian rocks with diverse appearances. The decoder is responsible for aggregating these features for pixelwise rock prediction. For evaluation, we establish two Mars rock datasets, including both real and synthesized images. One is MarsData-V2, an extension of our previously published MarsData collected from real Mars rocks. The other is SynMars, a synthetic dataset sequentially photographed from a virtual terrain built referring to the TianWen-1 dataset. Extensive experiments on the two datasets show the superiority of RockFormer for Martian rock segmentation, achieving state-of-the-art performance with decent computational simplicity.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Analysis of Lava Tubes’ Roughness and Radar Near-Nadir Regime
           Backscattering Properties

    • Free pre-print version: Loading...

      Authors: Leonardo Carrer;Lorenzo Bruzzone;
      Pages: 1 - 13
      Abstract: Lava tubes are terrestrial tunnel-like natural subsurface caves. Mounting evidence suggests their presence on the Moon and Mars. Planetary radar sounders are nadir-looking instruments operating in the high-frequency (HF)/very-HF (VHF) part of the spectrum with subsurface penetration capabilities. Recently, several studies either proposed future mission concepts for lava tubes’ detection or attempted to locate them on the Moon and Mars with the available radar-sounding data. Lava tubes are typically modeled as quasi-cylindrical structures but their actual geometry and their influence on the radar backscattering in near-nadir regime have never been investigated in the literature. These are crucial information for understanding the feasibility of detecting lava tubes by current and future planetary radar sounding systems. Accordingly, in this article: 1) we assess whether lava tubes are self-affine fractal surfaces at horizontal scales relevant to radio and microwave scattering and 2) we evaluate the effect of lava tube topography on the radar backscattering response in the near-nadir regime. Our experimental results, which are inferred from 3-D terrestrial laser scanning (TLS) data of planetary lava tube analogs, show that lava tubes: 1) are self-affine fractals at horizontal scales relevant to radar sounding and 2) they are electromagnetically rough surfaces, especially in the VHF band. We provide quantitative values on the lava tube fractal parameters and radar roughness losses along with a discussion on both: 1) the implication of our results on current radar sounding systems’ ability to detect lava tubes and 2) the planning of future missions devoted to lava tube detection and characterization.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Level-Aware Consistent Multilevel Map Translation From Satellite Imagery

    • Free pre-print version: Loading...

      Authors: Ying Fu;Zheng Fang;Linwei Chen;Tao Song;Defu Lin;
      Pages: 1 - 14
      Abstract: With the rapid development of remote sensing technology, the quality of satellite imagery (SI) is getting higher, which contains rich cartographic information that can be translated into maps. However, existing methods either only focus on generating single-level map or do not fully consider the challenges of multilevel translation from satellite imageries, i.e., the large domain gap, level-dependent content differences, and main content consistency. In this article, we propose a novel level-aware fusion network for the SI-based multilevel map generation (MLMG) task. It aims to tackle these three challenges. To deal with the large domain gap, we propose to generate maps in a coarse-to-fine way. To well-handle the level-dependent content differences, we design a level classifier to explore different levels of the map. Besides, we use a map element extractor to extract the major geographic element features from satellite imageries, which is helpful to keep the main content consistency. Next, we design a multilevel fusion generator to generate a consistent multilevel map from the multilevel preliminary map, which further ensures the main content consistency. In addition, we collect a high-quality multilevel dataset for SI-based MLMG. Experimental results show that the proposed method can provide substantial improvements over the state-of-the-art alternatives in terms of both objective metric and visual quality.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Easy-to-Hard Structure for Remote Sensing Scene Classification in
           Multitarget Domain Adaptation

    • Free pre-print version: Loading...

      Authors: Ba Hung Ngo;Yeon Jeong Chae;Jae Hyeon Park;Ju Hyun Kim;Sung In Cho;
      Pages: 1 - 15
      Abstract: Multitarget domain adaptation (MTDA) is a transfer learning task that uses knowledge extracted from a labeled source domain to adapt across multiple unlabeled target domains. The MTDA setting is more complicated than the single-source-single-target domain adaptation (S3TDA) setting because domain shift not only exists in each pair of a source–target domain but also exists among different target domains. In addition, multiple-target domains have their own unique characteristics because they are often collected from various conditions. The semantic information in each target domain can be damaged when they are naïvely merged into a single-target domain. Therefore, the trained model struggles to distinguish between representations in the combined target domain, which degrades the classification performance. Furthermore, the knowledge transferability from the source domain to multiple-target domains in prior studies leaves room for improvement because they only focus on exploiting the relationship of source–target pairs while failing to consider the correlation among multiple-target domains. This article introduces an easy-to-hard adaption structure to solve these problems in MTDA. The proposed method consists of three components: Extracting source representations, Hierarchical intratarget feature Alignment, and Collaborative intertarget feature Alignment, called EHACA. These components are used to encode the semantic information in each target domain and explore the relationships between the source and target domains, and among different target domains. The proposed method shows outstanding classification performance over five remote sensing datasets of MTDA tasks, surpassing state-of-the-art approaches in most experimental scenarios.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A CNN-Based Sentinel-2 Image Super-Resolution Method Using Multiobjective
           Training

    • Free pre-print version: Loading...

      Authors: Vlad Vasilescu;Mihai Datcu;Daniela Faur;
      Pages: 1 - 14
      Abstract: Deep learning methods have become ubiquitous tools in many Earth observation applications, delivering state-of-the-art results while proving to generalize for a variety of scenarios. One such domain concerns the Sentinel-2 (S2) satellite mission, which provides multispectral images in the form of 13 spectral bands, captured at three different spatial resolutions: 10, 20, and 60 m. This research aims to provide a super-resolution mechanism based on fully convolutional neural networks (CNNs) for upsampling the low-resolution (LR) spectral bands of S2 up to 10-m spatial resolution. Our approach is centered on attaining good performance with respect to two main properties: consistency and synthesis. While the synthesis evaluation, also known as Wald’s protocol, has spoken for the performance of almost all previously introduced methods, the consistency property has been overlooked as a viable evaluation procedure. Recently introduced techniques make use of sensor’s modulation transfer function (MTF) to learn an approximate inverse mapping from LR to high-resolution images, which is on a direct path for achieving a good consistency value. To this end, we propose a multiobjective loss for training our architectures, including an MTF-based mechanism, a direct input–output mapping using synthetically degraded data, along with direct similarity measures between high-frequency details from already available 10-m bands, and super-resolved images. Experiments indicate that our method is able to achieve a good tradeoff between consistency and synthesis properties, along with competitive visual quality results.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Accurate Label Refinement From Multiannotator of Remote Sensing Data

    • Free pre-print version: Loading...

      Authors: Xiangyu Wang;Lyuzhou Chen;Taiyu Ban;Derui Lyu;Yifeng Guan;Xingyu Wu;Xiren Zhou;Huanhuan Chen;
      Pages: 1 - 13
      Abstract: The remote sensing (RS) field has an increasing research interest in using deep learning (DL) models to recognize kinds of RS data, leading to a great demand for training data annotation. Due to the high cost of expertise, using nonexperts to label data has become an important way to improve labeling efficiency. Commonly, a single data sample is labeled by multiple annotators and the most voted label is accepted to promise accuracy. But in the RS context, the widely admitted strategy could lose effect. Usually RS data involve considerable classes on account of the complexity of surface environments, which is prone to interclass similarity difficult to distinguish. Annotators without expertise probably make mistakes on these indistinguishable classes, thus causing error voted labels. Although classification of different characteristics in RS data has been widely documented, the nonexpert annotators are unfamiliar with these expertise, and it is difficult to force them to handle specialized labeling skills. To address the issues, this article bases multiannotator label selection on the investigation of annotators’ own ability in distinguishing similar classes of images. A quality evaluation process is designed which weights the labels from capable annotators higher than those from weak ones. By a multi-round quality evaluation algorithm, correct labels could outcompete the wrong ones even disadvantaged in numbers. Experimental results demonstrate the advance of the proposed method on the RS datasets.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Universal Domain Adaptation for Remote Sensing Image Scene Classification

    • Free pre-print version: Loading...

      Authors: Qingsong Xu;Yilei Shi;Xin Yuan;Xiao Xiang Zhu;
      Pages: 1 - 15
      Abstract: The domain adaptation (DA) approaches available to date are usually not well suited for practical DA scenarios of remote sensing image classification since these methods (such as unsupervised DA) rely on rich prior knowledge about the relationship between label sets of source and target domains, and source data are often not accessible due to privacy or confidentiality issues. To this end, we propose a practical universal DA (UniDA) setting for remote sensing image scene classification that requires no prior knowledge on the label sets. Furthermore, a novel UniDA method without source data is proposed for cases when the source data are unavailable. The architecture of the model is divided into two parts: the source data generation stage and the model adaptation stage. The first stage estimates the conditional distribution of source data from the pretrained model using the knowledge of class separability in the source domain and then synthesizes the source data. With this synthetic source data in hand, it becomes a UniDA task to classify a target sample correctly if it belongs to any category in the source label set or mark it as “unknown” otherwise. In the second stage, a novel transferable weight that distinguishes the shared and private label sets in each domain promotes the adaptation in the automatically discovered shared label set and recognizes the “unknown” samples successfully. Empirical results show that the proposed model is effective and practical for remote sensing image scene classification, regardless of whether the source data are available or not. The code is available at https://github.com/zhu-xlab/UniDA.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Generative Building Feature Estimation From Satellite Images

    • Free pre-print version: Loading...

      Authors: Liu He;Jie Shan;Daniel Aliaga;
      Pages: 1 - 13
      Abstract: Urban and environmental researchers seek to obtain building features (e.g., building shapes, counts, and areas) at large scales. However, blurriness, occlusions, and noise from prevailing satellite images severely hinder the performance of image segmentation, super-resolution, or deep-learning-based translation networks. In this article, we combine globally available satellite images and spatial geometric feature datasets to create a generative modeling framework that enables obtaining significantly improved accuracy in per-building feature estimation and the generation of visually plausible building footprints. Our approach is a novel design that compensates for the degradation present in satellite images by using a novel deep network setup that includes segmentation, generative modeling, and adversarial learning for instance-level building features. Our method has proven its robustness through large-scale prototypical experiments covering heterogeneous scenarios from dense urban to sparse rural. Results show better quality over advanced segmentation networks for urban and environmental planning, and show promise for future continental-scale urban applications.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • TCD: Task-Collaborated Detector for Oriented Objects in Remote Sensing
           Images

    • Free pre-print version: Loading...

      Authors: Caiguang Zhang;Boli Xiong;Xiao Li;Gangyao Kuang;
      Pages: 1 - 14
      Abstract: Oriented object detection (OOD) in remote sensing image interpretation is challenging due to the difficulty of locating objects with arbitrary orientations. Existing methods have made considerable progress based on oriented heads or anchors. However, most of them follow the classical detection paradigm, such as assigning samples based on Intersection-over-Unions (IoU) and predicting through two independent tasks. These fixed strategies impair the consistency between classification and localization predictions, resulting in the prediction with optimal localization accuracy being suppressed by the nonoptimal ones during nonmaximum suppression (NMS). To address this problem, a task-collaborated detector (TCD) is proposed. Compared with current single-stage methods, its improvements include two aspects: task-collaborated assignment (TCA) and task-collaborated head (TCH). Specifically, to better pull closer the best anchors for two tasks, TCA introduces classification and localization confidence into sample assignment and tends to select the anchors with accurate and consistent predictions as positive during training. TCH provides a better balance for learning interactive and discriminative features. It can flexibly adjust the spatial feature distribution of classification and localization tasks by learning the joint features from the aggregation layer. Extensive experiments are conducted on HRSC2016, DOTA, and DIOR-R, and the proposed TCD achieves the state-of-the-art performance [90.60, 80.89, and 65.04 mean average precision (mAP), respectively]. Consistency analysis also demonstrates that TCD can significantly improve prediction consistency.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Physics Guided Remote Sensing Image Synthesis Network for Ship Detection

    • Free pre-print version: Loading...

      Authors: Weichang Zhang;Rui Zhang;Guoqing Wang;Wei Li;Xun Liu;Yang Yang;Die Hu;
      Pages: 1 - 14
      Abstract: Automatic detection and localization of objects in remote sensing images are of great significance for remote sensing systems. Existing frameworks usually train an object detection network using collected remote sensing images. However, these models usually perform poorly due to the lack of large-scale training datasets, which is often the case for special remote sensing scenarios, e.g., the detection of ships in the open sea. Although image synthesis is a common strategy to alleviate the issue of data insufficiency, the trained model still performs poorly when being tested on real-world scenes. Aimed at this, a novel sensor-related image synthesis framework, dubbed as remote sensing-image synthesis pipeline (RS-ISP), is developed to address the lack of on-orbit remote sensing images. Specifically, our RS-ISP introduces two novel designs to ensure the distribution consistency between the generated images and the real images: 1) the first is a novel pipeline for modeling the physical process of noise production during image capture using specific sensors and 2) the second is the design of a detection-oriented image harmonization model. Similar to the existing design, our model first produces coarse synthetic images by copy–paste operation, on which the proposed harmonization process is used to reduce the variation in the pasted foreground and background. By incorporating these two designs into a unified framework, our RS-ISP is designed and used to produce large-scale synthetic images used to train the object detection model for detecting ships in remote sensing images. Comparative experiments demonstrated that RS-ISP increased the AP@.50 from 0.148 to 0.498 for the ship detection task. Code will be publicly available.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Deep Saliency Smoothing Hashing for Drone Image Retrieval

    • Free pre-print version: Loading...

      Authors: Yaxiong Chen;Jinghao Huang;Lichao Mou;Pu Jin;Shengwu Xiong;Xiao Xiang Zhu;
      Pages: 1 - 13
      Abstract: Deep hashing algorithms are widely exploited in retrieval tasks due to their low storage and retrieval efficiency. Most of them focus on global feature learning, while neglecting local fine-grained features and saliency information for drone images. In this article, we tackle these dilemmas with a novel deep saliency smoothing hashing (DSSH) algorithm, which can leverage saliency capture mechanism, distribution smoothing term, global features, and local fine-grained features to learn effective hash codes for drone image retrieval. The DSSH algorithm first designs an information extraction module to capture global features and local fine-grained features for drone images. Meanwhile, a saliency capture module is proposed to perform information interaction attention and visual enhancement attention, which can capture the saliency area of drone images effectively. On top of the two paths, a novel objective function is designed to preserve the similarity of hash codes, smooth the distribution of drone image datasets, and reduce the quantization errors between hash codes and hash-like codes concurrently. Extensive experiments on the Drone Action Dataset and ERA Drone Dataset demonstrate that the DSSH algorithm can further improve the retrieval performance compared with other deep hashing algorithms.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Evaluation of Three Land Surface Temperature Products From Landsat Series
           Using in Situ Measurements

    • Free pre-print version: Loading...

      Authors: Mengmeng Wang;Can He;Zhengjia Zhang;Tian Hu;Si-Bo Duan;Kaniska Mallick;Hua Li;Xiuguo Liu;
      Pages: 1 - 19
      Abstract: Three operational long-term land surface temperature (LST) products from Landsat series are available to the community until now, i.e., U.S. Geological Survey (USGS) LST, Instituto Português do Mar e da Atmosfera (IPMA) LST, and China University of Geosciences (CUG) LST. A comprehensive assessment of these LST products is essential for their subsequent applications (APPs) in energy, water, and carbon cycle modeling. In this study, an evaluation of these three Landsat LST products was performed using in situ LST measurements from five networks [surface radiation budget (SURFRAD), atmospheric radiation measurement (ARM), Heihe watershed allied telemetry experimental research (HiWATER), baseline surface radiation network (BSRN), and National Data Buoy Center (NDBC)] for the period of 2009–2019. Results reveal that the overall accuracies of CUG LST with bias [root-mean-square error (RMSE)] of 0.54 K (2.19 K) and IPMA LST with bias (RMSE) of 0.59 K (2.34 K) are marginally superior to USGS LST with bias (RMSE) of 0.96 K (2.51 K). The RMSE of USGS LST is about 0.3 K less than IPMA/CUG LST at water surface sites and is about 0.4 K higher than IPMA/CUG LST at cropland and shrubland sites. As for tundra, grassland, and forest sites, the RMSEs of three Landsat LST products are similar, and the RMSE difference among three Landsat LST products is < 0.18 K. Considering the close emissivity estimates over water surface in these three LST data, USGS LST has a better performance in atmospheric correction over water surface compared with IPMA/CUG LST. For land surface sites, the RMSE of LST increases initially and then decreases with land surface emissivity (LSE) for three Landsat LST products. This indicates that the emissivity correction has a large uncertainty for moderately vegetated surface with emissivity ranging from 0.970 to 0.980. Underestimated emissivity for USGS LST at vegetated sites leads to overestimation of LST- which could have led to the higher bias and RMSE compared with IPMA/CUG LST. For the LST retrievals for the three different sensors [i.e., Thematic Mapper (TM), Enhanced Thematic Mapper Plus (ETM+), and thermal infrared sensor (TIRS)] onboard the Landsat satellite series, the accuracies are consistent and comparable, which is beneficial for providing long-term and coherent LST.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • KCPNet: Knowledge-Driven Context Perception Networks for Ship Detection in
           Infrared Imagery

    • Free pre-print version: Loading...

      Authors: Yaqi Han;Jingwen Liao;Tianshu Lu;Tian Pu;Zhenming Peng;
      Pages: 1 - 19
      Abstract: Ship detection plays a crucial role in a variety of military and civilian marine inspection applications. Infrared images are irreplaceable data sources for ship detection due to their strong adaptability and excellent all-weather reconnaissance ability. However, previous researches mainly focus on visible light or synthetic aperture radar (SAR) ship detection, while infrared ship detection is left in a huge blind spot. The main obstacles to this dilemma lie in the absence of public datasets, small scale, and poor semantic information of infrared ships, and severe clutter in complex ocean environments. To address the above challenges, we propose a knowledge-driven context perception network (KCPNet) and construct a public dataset called infrared ship detection dataset (ISDD). In KCPNet, aiming at the small scale of infrared ships, a balanced feature fusion network (BFF-Net) is proposed to balance information from all backbone layers and generate nonlocal features with balanced receptive fields. Moreover, considering the key role of contextual information, a contextual attention network (CA-Net) is designed to improve robustness in complex scenes by enhancing target and contextual information and suppressing clutter. Inspired by prior knowledge of human cognitive processes, we construct a novel knowledge-driven prediction head to autonomously learn visual features and back-propagate the knowledge throughout the whole network, which can efficiently reduce false alarms. Extensive experiments demonstrate that the proposed KCPNet achieves state-of-the-art performance on ISDD. Source codes and ISDD are accessible at https://github.com/yaqihan-9898.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Thermal UAV Image Super-Resolution Guided by Multiple Visible Cues

    • Free pre-print version: Loading...

      Authors: Zhicheng Zhao;Yong Zhang;Chenglong Li;Yun Xiao;Jin Tang;
      Pages: 1 - 14
      Abstract: Unmanned aerial vehicle (UAV) thermal-imaging has received much attention, but the insufficient image resolution caused by thermal imaging systems is still a crucial problem that limits the understanding of thermal UAV images. However, high-resolution visible images are relatively easy to access, and it is thus valuable for exploring useful information from visible image to assist thermal UAV image super-resolution (SR). In this article, we propose a novel multiconditioned guidance network (MGNet) to effectively mine the information of visible images for thermal UAV image SR. High-resolution visible UAV images usually contain salient appearance, semantic, and edge information, which plays a critical role in boosting the performance of thermal UAV image SR. Therefore, we design an effective multicue guidance module (MGM) to leverage the appearance, edge, and semantic cues from visible images to guide thermal UAV image SR. In addition, we build the first benchmark dataset for the task of thermal UAV image SR guided by visible images. It is collected by a multimodal UAV platform and composes of 1025 pairs of manually aligned visible and thermal images. Extensive experiments on the built dataset show that our MGNet can effectively leverage useful information from visible images to improve the performance of thermal UAV image SR and perform well against several state-of-the-art methods. The dataset is available at: https://github.com/mmic-lcl/Datasets-and-benchmark-code.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Sparse Regularization-Based Spatial–Temporal Twist Tensor Model for
           Infrared Small Target Detection

    • Free pre-print version: Loading...

      Authors: Jie Li;Ping Zhang;Lingyi Zhang;Zhiyuan Zhang;
      Pages: 1 - 17
      Abstract: Infrared (IR) small target detection under complex environments is an essential part of IR search and track systems. However, previously proposed IR small target detection algorithms cannot achieve complete suppression of complex and significant backgrounds. The spatial–temporal information of image sequences is not fully exploited. In this article, we present a sparse regularization-based twist tensor model for IR small target detection. First, the twist tensor model is built via perspective conversion based on the target’s local continuity in the spatial–temporal domain, which makes the original complicated background components more structured and increases the difference between the background and the target. Then, the structured sparsity-inducing norm is introduced to define the locality and continuity of the target. To further minimize the sparse background structures and global noise, the structured sparsity-inducing norm and the $l_{1}$ norm are combined as the target’s parse constraint. Experimental results on real scenes reveal that the suggested method can process images with high detection accuracy and outstanding background suppression ability compared to various state-of-the-art methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Receptive-Field and Direction Induced Attention Network for Infrared Dim
           Small Target Detection With a Large-Scale Dataset IRDST

    • Free pre-print version: Loading...

      Authors: Heng Sun;Junxiang Bai;Fan Yang;Xiangzhi Bai;
      Pages: 1 - 13
      Abstract: Infrared small target detection plays an important role in military and civilian fields while it is difficult to be solved by deep learning (DL) technologies due to scarcity of data and strong interclass imbalance. To relieve scarcity of data, we build a massive dataset IRDST, which contains 142 727 frames. Also, we propose a receptive-field and direction-induced attention network (RDIAN), which is designed using the characteristics of target size and grayscale to solve the interclass imbalance between targets and background. Using convolutional layers with different receptive fields in feature extraction, target features in different local regions are captured, which enhances the diversity of target features. Using multidirection guided attention mechanism, targets are enhanced in low-level feature maps. Experimental results with comparison methods and ablation study demonstrate effective detection performance of our model. Dataset and code will be available at https://xzbai.buaa.edu.cn/datasets.html.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Optimizing the Protocol of Near-Surface Remote Sensing Experiments Over
           Heterogeneous Canopy Using DART Simulated Images

    • Free pre-print version: Loading...

      Authors: Biao Cao;Jean-Philippe Gastellu-Etchegorry;Tiangang Yin;Zunjian Bian;Junhua Bai;Junyong Fang;Boxiong Qin;Yongming Du;Hua Li;Qing Xiao;Qinhuo Liu;
      Pages: 1 - 16
      Abstract: Optical canopy models that connect land surface properties and satellite-observed radiance must be validated before being used. These models include the bidirectional reflectance distribution function (BRDF) models in the visible and near-infrared domains, and directional brightness temperature (DBT) models in the thermal infrared domain. Near-surface experiments have been extensively conducted to evaluate the modeling accuracy, including ground-, tower-, and aircraft-based measurements. Indeed, it should be noted that in situ measured BRDF/DBT results are sensitive to the experiment protocol, such as sensor moving orientation, flight height, and sampling frequency. A practical tool for optimizing the in situ measurement protocols is needed in the community of remote sensing modeling. For that, we devised a virtual experiment framework based on the discrete anisotropic radiative transfer (DART) 3-D radiative transfer model that is capable of simultaneously simulating both the BRDF/DBT pattern and the images acquired by in situ cameras. Here, as an optimization case, we use it to determine the optimal sensor flight orientation over heterogeneous vegetated canopies (a row-planted scene with three solar angles and a discrete scene with three solar angles) for measuring their DBT distribution. Results showed considerable errors (i.e., image-extracted DBT minus DART-simulated DBT) exist for sensor flight orientation along the canopy rows ( $R^{2}$ = 0.24 and root mean square error (RMSE) = 4.32 K), and they become much smaller ( $R^{2}$ = 0.94 ~ 0.98 and RMSE = 0.82 ~ 1.03 K) in other typical orientations (e.g., cross row plane, solar principal plane, and cross solar principal plane). The critical azimuth offset rela-ive to the row direction that can ensure an acceptable RMSE < 1 K is quantified as atan(3*Unitwidth/Scenesize) based on a series of intensive simulations by this new tool. However, the RMSE of the discrete scene is not sensitive to the flight orientation. Such accuracy differences in various protocols were experimentally verified over row-planted maize using a 4-D tower in Huailai, Hebei, China. The result highlights the great potential of this newly designed DART-based virtual experiment to optimize near-surface experiment protocols.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Detection of Subtle Thermal Anomalies: Deep Learning Applied to the ASTER
           Global Volcano Dataset

    • Free pre-print version: Loading...

      Authors: Claudia Corradino;Michael S. Ramsey;Sophie Pailot-Bonnétat;Andrew J. L. Harris;Ciro Del Negro;
      Pages: 1 - 15
      Abstract: Twenty-one years of Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) global thermal infrared (TIR) acquisitions provide a large amount of data for volcano monitoring. These data, with high spatial and spectral resolution, enable routine investigations of volcanoes in remote and inaccessible regions, including those with no ground-based monitoring. However, the dataset is too large to be manually analyzed on a global basis. Here, we systematically process the data over several volcanoes using a deep learning algorithm to automatically extract volcanic thermal anomalies. We explore the application of a convolutional neural network (CNN), specifically UNET, to detect subtle to intense anomalies exploiting the spatial relationships of the volcanic features. We employ a supervised UNET network trained with the largest (1500) labeled dataset of ASTER TIR images from five different volcanoes, namely, Etna (Italy), Popocatépetl (Mexico), Lascar (Chile), Fuego (Guatemala), and Kliuchevskoi (Russia). We show that our approach achieves high accuracy (93%) with excellent generalization capabilities. The effectiveness of our model for detecting the full range of thermal emission is shown for volcanoes with very different styles of activity and tested at Vulcano (Italy). The results demonstrate the potential applicability of the proposed approach to the development of automated thermal analysis systems at the global scale using future TIR data such as the planned NASA Surface Biology and Geology (SBG) mission.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Infrared Small Target Detection Based on Local Contrast-Weighted
           Multidirectional Derivative

    • Free pre-print version: Loading...

      Authors: Yunkai Xu;Minjie Wan;Xiaojie Zhang;Jian Wu;Yili Chen;Qian Chen;Guohua Gu;
      Pages: 1 - 16
      Abstract: Realizing robust infrared small target detection in complex backgrounds is of great essence for infrared search and tracking (IRST) applications. However, the high-intensity structures in background regions, such as the sharp edges, make it a challenging task, especially when the target is with low signal-to-clutter ratio (SCR). To address this issue, we propose an infrared small target detection method using local contrast-weighted multidirectional derivative (LCWMD). It is a robust detector that comprehensively considers the target property, background information, and the relation between them. First, we consider the approximate isotropy of the infrared small target and present a new multidirectional derivative with penalty factors based on the Facet model to develop the target salience in the local region. Second, a dual local contrast fusion model with the trilayer design is introduced to amplify the difference between the target and the background, so as to further suppress the high-intensity structural clutters. Finally, the LCWMD map is obtained by weighting the above two filtered maps, after which an adaptive segmentation operation is applied to accomplish the target detection. The results of comparative experiments implemented on real infrared images demonstrate that our method outperforms other state-of-the-art detectors by several times in terms of SCR gain (SCRG) and background suppression factor (BSF).
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • One-Stage Cascade Refinement Networks for Infrared Small Target Detection

    • Free pre-print version: Loading...

      Authors: Yimian Dai;Xiang Li;Fei Zhou;Yulei Qian;Yaohong Chen;Jian Yang;
      Pages: 1 - 17
      Abstract: Single-frame infrared small target (SIRST) detection has been a challenging task due to a lack of inherent characteristics, imprecise bounding box regression, a scarcity of real-world datasets, and sensitive localization evaluation. In this article, we propose a comprehensive solution to these challenges. First, we find that the existing anchor-free label assignment method is prone to mislabeling small targets as background, leading to their omission by detectors. To overcome this issue, we propose an all-scale pseudobox-based label assignment scheme that relaxes the constraints on the scale and decouples the spatial assignment from the size of the ground-truth target. Second, motivated by the structured prior of feature pyramids, we introduce the one-stage cascade refinement network (OSCAR), which uses the high-level head as soft proposal for the low-level refinement head. This allows OSCAR to process the same target in a cascade coarse-to-fine manner. Finally, we present a new research benchmark for infrared small target detection, consisting of the SIRST-V2 dataset of real-world, high-resolution single-frame targets, the normalized contrast evaluation metric, and the DeepInfrared toolkit for detection. We conduct extensive ablation studies to evaluate the components of OSCAR and compare its performance to state-of-the-art model- and data-driven methods on the SIRST-V2 benchmark. Our results demonstrate that a top-down cascade refinement framework can improve the accuracy of infrared small target detection without sacrificing efficiency. The DeepInfrared toolkit, dataset, and trained models are available at https://github.com/YimianDai/open-deepinfrared.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Physical-Based Method for Pixel-by-Pixel Quantifying Uncertainty of Land
           Surface Temperature Retrieval From Satellite Thermal Infrared Data Using
           the Generalized Split-Window Algorithm

    • Free pre-print version: Loading...

      Authors: Yang Gui;Si-Bo Duan;Zhao-Liang Li;Cheng Huang;Meng Liu;Xiangyang Liu;Caixia Gao;
      Pages: 1 - 15
      Abstract: Land surface temperature (LST) is an important physical parameter at the interface between the Earth’s surface and the atmosphere. Accurately quantifying LST uncertainty is essential for the generation of a long-term and consistent LST Climate Data Record (CDR) or Earth System Data Record (ESDR) from either multiple sensors or algorithms. In this study, a physical-based method was proposed to quantify the uncertainty of LST retrieval from satellite thermal infrared (TIR) data using the generalized split-window (GSW) algorithm. LST uncertainties were parameterized as a function of brightness temperature at the top of the atmosphere (TOA) and surface emissivity in two split-window channels, which are two key input parameters in the GSW algorithm, as well as their uncertainties. The performance of the parameterized uncertainty model was evaluated according to the simulation dataset at six prescribed viewing zenith angles (VZAs) of 0°, 33.56°, 44.42°, 51.32°, 56.25°, and 60°, with a root mean squared error (RMSE) of 0.001 K. The coefficients of the parameterized uncertainty model at arbitrary VZA within a sensor’s field of view (FOV) can be obtained by linear interpolation of the coefficients at the six prescribed VZAs. Once the coefficients of the parameterized uncertainty model for each pixel are available, total LST uncertainties can be quantified on a pixel-by-pixel basis. As an example, the parameterized uncertainty model was applied to actual MODIS data for displaying the spatial distribution of LST uncertainties. The results indicate that the parameterized uncertainty model can characterize the spatial variation in LST uncertainties well over various land cover types.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Chang’E-4 Measurements of Lunar Surface Temperatures: Thermal
           Conductivity of the Near Surface Regolith

    • Free pre-print version: Loading...

      Authors: Wenchao Zheng;Guoping Hu;Yunzhao Wu;Qi Jin;Zhengmei Li;Li Feng;
      Pages: 1 - 14
      Abstract: Lunar thermal conductivity is significantly important for understanding the geological processes of the Moon. The Apollo in situ heat flow experiments and diviner remote sensing data provide us with good constraints on the thermophysical properties of the lunar regolith fines layer. As the first farside in situ experiments, Chang’E-4 (CE-4) temperature sensors will further extend our understanding. In this study, a 1-D thermal model with updated bulk densities by the CE-4 lunar penetrating radar (LPR) data is used. Then, the nighttime surface temperatures from the sensors of the CE-4 lander are applied to retrieve the surface thermal conductivity at the CE-4 landing site. After minimizing the differences between the predicted temperature and CE-4 measurements, the range of surface layer contact conductivity ( $k_{s}$ ) within the upper 1 cm is about (0.95–2.26) $times 10^{-3},,text {Wm}^{-1}cdot text {K}^{-1}$ for T1, (1.04–2.44) $times 10^{-3},,text {Wm}^{-1}cdot text {K}^{-1}$ for T2, (0.60–1.40) $times 10^{-3},,text {Wm}^{-1}cdot text {K}^{-1}$ for T3, and (0.60–1.39) $times 10^{-3},,text {Wm}^{-1}cdot text {K}^{-1}$ for T4, respectively. In addition, the factors that may affect the inversion or the measurements were discussed, such as the shadowing effect, the rock abundance (RA), the lateral conduction from the rover transfer mechanism, and the densification effect during the measurements.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Temporal Upscaling of MODIS 1-km Instantaneous Land Surface Temperature to
           Monthly Mean Value: Method Evaluation and Product Generation

    • Free pre-print version: Loading...

      Authors: Xiangyang Liu;Zhao-Liang Li;Jia-Hao Li;Pei Leng;Meng Liu;Maofang Gao;
      Pages: 1 - 14
      Abstract: The monthly mean land surface temperature (MMLST) reflects more stable intra- and interannual temperature variations, and therefore, it has a wider range of applications than instantaneous land surface temperature (LST). This study aimed to generate a high-resolution global MMLST product by temporally upscaling the Moderate Resolution Imaging Spectroradiometer (MODIS) 1-km instantaneous LST. First, six current methods were comprehensively evaluated using cross-validation technology. These six methods are the cross combinations of two temporal aggregation schemes: the average by observations (ABO) and average by days (ABD), and three conversion models: the diurnal temperature cycle model (DTC), the simple average of two instantaneous LSTs (TSA), and a weighted average model for multiple instantaneous LSTs (MWA). The analysis with measurements from 235 flux stations worldwide revealed that the choice of conversion model considerably affected the overall retrieval accuracy, whereas the influence of the aggregation scheme was minor. From the conversion model standpoint, MWA performed best, followed by DTC, and finally TSA; this order remained the same even if DTC and TSA were improved with mean bias correction. Notably, the errors of ABDMWA decreased as the number of daily mean LST (NOD) increased, whereas the errors of ABOMWA were not related to NOD. Accordingly, we deduced that the optimal strategy for estimating MMLST is using ABOMWA when NOD is < 20 and ABDMWA when NOD is $ge 20$ . Subsequently, we adopted this combination method to process MODIS instantaneous LSTs and produced a global 1-km MMLST dataset for the years 2003–2020. The validation showed a satisfactory accuracy with a root mean square error (RMSE) of 1.6 K. The intercomparison with MMLSTs from geostationary (GEO) satellites (containing complete LST daily cycle) presented a good ag-eement (biases < 0.3 K and STDs < 2 K). Compared with atmospheric infrared sounder (AIRS) L3 monthly standard physical retrieval (AIRS3STM) product which had the same temporal span, the newly generated product exhibited a high consistency in reflecting temporal variations of global temperature. Most importantly, it had a prominently better ability to retrieve spatial details of temperature variations due to its higher resolution. Our new method and product show promising prospects for applications in global change studies, where accurate spatially resolved MMLST data are one of the fundamental geophysical variables required.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Novel Two-Stage Destriping Algorithm Based on MWIR Energy Separation and
           Image Guidance (MES-IG)

    • Free pre-print version: Loading...

      Authors: Xinyu Zhou;Ye Zhang;Jinhao Liu;Yue Hu;
      Pages: 1 - 16
      Abstract: Long-wave infrared (LWIR) bands in multispectral datasets are extremely useful in many applications. However, the LWIR bands usually suffer from undesirable stripe noise, which impedes their further application. Compared with emission-dominated LWIR, the mid-wave infrared (MWIR) bands containing both emitted and reflected radiation usually exhibit higher image quality. In this article, we propose a novel two-stage MWIR energy separation and image guidance (MES-IG) algorithm to destripe the LWIR images with the assistance of the MWIR bands. In the first stage, we decompose the MWIR image into the emitted and reflected components by solving a constrained optimization problem. Specifically, we impose the low-rank penalty to enforce the similarities between MWIR and LWIR, and we use the total variation (TV) regularization to exploit the similarities between MWIR and visible and near-infrared (VNIR) images. In the second stage, the obtained emitted component of MWIR is considered as the guidance image to remove the stripes in the LWIR images by adopting the 1-D guided filter algorithm. Numerical experiments on the Chinese Gaofen-5 satellite and the Moderate Resolution Imaging Spectroradiometer (MODIS) data demonstrate the utility of the proposed method in providing improved LWIR image destriping performance over the state-of-the-art algorithms.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Analytical Study of the Changes in Brightness Temperature Based on the
           Tectonic Field Associated With Three Earthquakes in the Eastern Tibetan
           Plateau

    • Free pre-print version: Loading...

      Authors: Xing Yang;Qian Lu;Tiebao Zhang;Fang Du;Feng Long;Min Zhao;Xiaofeng Liao;
      Pages: 1 - 15
      Abstract: The thermal infrared brightness temperature (BT) of the eastern Tibetan Plateau (TP) was retrieved from the Moderate Resolution Imaging Spectroradiometer (MODIS) level-1B data. The multiyear averaged BT background field was subtracted from the punctual BT data to yield monthly BT spatial anomaly, and calculated time series of BT for the secondary blocks. Then, the spatial and temporal changes in the BT of the study area before the Menyuan M6.4, Zaduo M6.2, and Jiuzhaigou M7.0 earthquakes were investigated and analyzed based on the tectonic setting. The results show the following. The spatial BT radiation enhancement frequency rose remarkably before strong earthquakes; each of the three earthquakes was preceded by marked spatiotemporal continuous BT anomalies. The tectonic setting significantly influences the BT anomaly feature. The spatial BT anomaly was not notable in the Qaidam and Qilian block before the Menyuan earthquake; the spatial BT anomaly mainly appeared in the Qiangtang and Bayan Har blocks before the Zaduo and Jiuzhaigou earthquakes. The Qiangtang and Bayan Har block’s BT time series curves have similar features. The Qaidam and Qilian block’s BT time series curves have analogous shapes. The three earthquakes may be regarded as one seismic event induced by a stage of tectonic stress enhancement rather than three independent occasions. The spatial BT anomalous behavior before earthquakes is, to a great extent, like the result of the rock stress loading experiment; the rock compression and the lithosphere–atmosphere–ionosphere coupling (LAIC) may be the main reasons for the intensification of the BT radiation.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Estimating Hourly All-Weather Land Surface Temperature From FY-4A/AGRI
           Imagery Using the Surface Energy Balance Theory

    • Free pre-print version: Loading...

      Authors: Weihan Liu;Jie Cheng;Qiao Wang;
      Pages: 1 - 18
      Abstract: Thermal infrared (TIR) observations from geostationary satellites enable the retrieval of the diurnal variations in land surface temperatures (LSTs) linked to the land–atmosphere energy exchange and the water cycle. However, cloud cover obstructs TIR signals and leads to missing gaps that generally account for more than 50% of TIR LST maps. This study proposed an effective method to estimate hourly all-weather LST from the Advanced Geosynchronous Radiation Imager (AGRI) data under the framework of surface energy balance (SEB) theory. An improved temperature and emissivity separation algorithm was first used to obtain the high-quality clear-sky LST, which plays a decisive role in ensuring the accuracy of recovered cloudy-sky LST. Then, we proposed a unique way to solve the temperature difference ( $Delta $ LST) between the cloudy-sky LST and hypothetical clear-sky LST caused by cloud radiative effects (CREs). The bias and root mean squared error (RMSE) of the estimated AGRI hourly cloudy-sky LST are 0.10 K and 3.71 K during the daytime, and −0.20 K and 2.73 K during the nighttime, respectively. The overall bias (RMSE) of the estimated AGRI all-weather LST is 0.02 K (2.84 K). The estimated hourly all-weather LST not only captures the rapid variation in diurnal LST but is also promising for temporal upscaling. The temporally upscaled daily mean LST shows a bias (RMSE) of 0.03 K (1.35 K). This study provides a promising solution to generate diurnal hourly all-weather LST for AGRI and other geostationary satellites.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Discrimination Between Dry and Water Ices by Full Polarimetric Radar:
           Implications for China’s First Martian Exploration

    • Free pre-print version: Loading...

      Authors: Hai Liu;Jianhui Li;Xu Meng;Bin Zhou;Guangyou Fang;Billie F. Spencer;
      Pages: 1 - 11
      Abstract: China’s first Mars rover named “Zhurong” has begun its exploration at the surface of the Utopian Plain in the Martian northern hemisphere on May 22, 2021. Mars rover penetrating radar (RoPeR) is one of the key payloads onboard the Zhurong rover and one of its prima scientific objectives is to detect potential water ice and/or dry ice in the subsurface soil at the landing site. The high-frequency channel of RoPeR, which is equipped with a fully-polarized antenna array with a center frequency of 1.3 GHz, can record full polarimetric radar reflections from subsurface anomalies. In this article, a radar system with a polarimetric antenna array the same as RoPeR was set up and carefully calibrated. Laboratory experiments were carried out to test the feasibility of RoPeR in the detection and discrimination of potential dry ice and/or water ice in Martian soils. The experimental results indicate that the reflection signals from the bottom of the dry ice and water ice samples present different polarimetric scattering characterizations. The characteristic polarimetric scattering features of dry ice and water ice revealed in this article would guide the analysis and interpretation of the real data acquired by RoPeR on Mars.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Joint Design Methods of Unimodular Sequences and Receiving Filters With
           Good Correlation Properties and Doppler Tolerance

    • Free pre-print version: Loading...

      Authors: Fulai Wang;Xiang-Gen Xia;Chen Pang;Xu Cheng;Yongzhen Li;Xuesong Wang;
      Pages: 1 - 14
      Abstract: Sequence set design with good correlation properties and Doppler tolerance has been a classic and important problem in many multichannel systems, including but not limited to the multiple-input multiple-output (MIMO) and simultaneous polarimetric radar systems. In this article, we first consider the problem of jointly designing Doppler resilient unimodular sequences and receiving filters to minimize the metric weighted integrated sidelobe level (WISL), which can be used to construct sequence sets with “thumbtack” co-channel and zero cross-channel ambiguity functions (AFs). To control the signal-to-noise ratio (SNR) loss caused by the mismatched filter, a peak constraint function is added to the objective function based on the penalty function method. An algorithm based on the alternatively iterative scheme and general majorization-minimization (MM) method is developed to tackle the constrained joint design problem. Moreover, the proposed algorithm is then extended to optimize the ${l_{p}}$ -norm of sidelobes of AFs, which gives a way to optimize the weighted peak sidelobe level (WPSL) metric. Due to the use of the fast Fourier transform (FFT) algorithm and a general acceleration scheme, the proposed algorithm can be realized efficiently. A number of simulations are provided to demonstrate the excellent performance of the proposed sequence set synthesis algorithms. Besides, an application of using the designed Doppler resilient sequence set on the simultaneous polarimetric radar to detect multiple moving targets in a strong clutter scene is given via simulation, which further verifies the practicability of the proposed algorithm.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • DOD and DOA Estimation From Incomplete Data Based on PARAFAC and Atomic
           Norm Minimization Method

    • Free pre-print version: Loading...

      Authors: Sizhe Gao;Hui Ma;Hongwei Liu;Yang Yang;
      Pages: 1 - 14
      Abstract: In this article, we propose an efficient direction of departure (DOD) and direction of arrival (DOA) estimation method for bistatic multiple-input multiple-output (MIMO) radar with faulty arrays. A third-order tensor model is built, and the measurement 3-D structure can be better utilized than the traditional matrix model. Subsequently, the atomic norm minimization (ANM) technique is used to further improve the angle estimation performance. Furthermore, we found in the research process that when the faulty arrays still maintain the symmetry property, the measurement tensor can be converted to the real-valued domain by the forward–backward averaging technique and the unitary transform technique. The new algorithm we proposed exploits the multidimensional structure of the signal without estimating the signal subspace. Comparing with traditional matrix completion (MC) methods, it has a better performance in terms of robustness and resolving correlated targets. Also, the algorithm proposed in this article does not require angle pairing. Simulation results verify the effectiveness of the proposed algorithm.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Range-Spread Target Detection Based on Adaptive Scattering Centers
           Estimation

    • Free pre-print version: Loading...

      Authors: Zhouchang Ren;Wei Yi;Wenjing Zhao;Lingjiang Kong;
      Pages: 1 - 14
      Abstract: Proper prior knowledge of target scattering centers (SCs) can help to obtain better detection performance of range-spread targets. However, target SCs are sensitive to the target’s attitude relative to the radar and vary significantly among different targets. The existing approaches that employ predetermined prior knowledge may suffer performance degradation when the prior information does not match the practical scenarios. A possible way to circumvent this drawback is to estimate the SCs of different targets adaptively and check the presence of a target utilizing the range cells occupied by the most likely target SCs. For this reason, this article develops a generalized likelihood ratio test based on adaptive SCs estimation (ASCE-GLRT) for range-spread target detection in compound-Gaussian clutter. Under the assumption that the target SCs are sparse, we model the problem of SCs estimation as a sparse signal representation. Moreover, since the sparse assumption may not always be satisfied in practice, a modified sparsity regularization method is proposed to enhance the robustness of the estimation performance of targets with different scattering characteristics. A theoretical analysis shows that the proposed detector can achieve the constant false alarm rate (CFAR) property. The performance assessments conducted by numerical simulation and field tests confirm the effectiveness and robustness of the proposed detector.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multiple Objects Automatic Detection of GPR Data Based on the AC-EWV and
           Genetic Algorithm

    • Free pre-print version: Loading...

      Authors: Guangyan Cui;Jie Xu;Yanhui Wang;Shengsheng Zhao;
      Pages: 1 - 16
      Abstract: The automatic detection of multiple objects in ground penetrating radar (GPR) data is investigated by searching for the reflected hyperbolas of buried objects, which can reduce the subjectivity of operators and improve identification accuracy. Based on Frequency-wavenumber (F-K) migration, the accurate calculation of electromagnetic wave velocity (AC-EWV) is proposed by searching for the minimum image entropy of migrated radargrams. To avoid global searching, potential positions of object hyperbolas are selected from the binarized radargram through the vertical gray gradient searching, then the sub_window is extracted with the potential position as the center. The best fitting hyperbola is detected with the genetic algorithm (GA) in the sub_window, and objects are finally determined with five hyperbolic matching criteria and the auto-categorization. This technique is verified with the simulated and measured GPR data about rebars, pipelines, and voids, and results demonstrate that it achieves the average correct rate, average missed rate, and the average misjudged rate is 98.46%, 1.33%, and 0.36%, respectively, and the average correct rate for GPR data of the double-layer rebars is 91.67%.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Space Target Classification With Corrupted HRRP Sequences Based on
           Temporal–Spatial Feature Aggregation Network

    • Free pre-print version: Loading...

      Authors: Yuan-Peng Zhang;Lei Zhang;Le Kang;Huan Wang;Ying Luo;Qun Zhang;
      Pages: 1 - 18
      Abstract: High-resolution range profile (HRRP) sequences have great potential for space target classification because they can provide both scattering information and micromotion information. However, many factors cause an obtained HRRP sequence for a space target to be corrupted in real cases due to noise interference, limited radar resources, and the requirement of multitarget observations. Many space target classification methods cease to be effective when HRRP sequences are corrupted, so classifying space targets with corrupted HRRP sequences is still a challenging problem. To solve this problem, a novel space target classification method based on a temporal–spatial feature aggregation network (TSFA-Net) is proposed by using the corrupted HRRP sequences directly. First, a sequence-to-token module (S2T-module) is designed to extract low-level and fine-grained features from the raw inputs. Second, to effectively model the long-range dependencies among corrupted HRRP sequences and capture global representations without losing target local features, we propose a parallel and dual-branch block, i.e., a temporal–spatial feature aggregation block (TSFA-block), by combining a Transformer network and a convolutional neural network (CNN). Then, via progressively hierarchically stacking TSFA-blocks, a hierarchical temporal–spatial feature aggregation subnetwork (H-TSFA-subnetwork) is constructed to obtain the final temporal–spatial features. Finally, a token-to-label module (T2L-module) is adopted to obtain the classification results. Extensive experiments demonstrate that the proposed method achieves state-of-the-art classification accuracy for space target classification with HRRP sequences, especially under the conditions of a low signal-to-noise ratio and a high missing rate.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Retrieval of Sea Surface Currents and Directional Wave Spectra by 24 GHz
           FMCW MIMO Radar

    • Free pre-print version: Loading...

      Authors: Giovanni Ludeno;Ilaria Catapano;Francesco Soldovieri;Gianluca Gennarelli;
      Pages: 1 - 13
      Abstract: This article investigates the capabilities of 24 GHz frequency-modulated continuous-wave (FMCW) multiple-input–multiple-output (MIMO) radar technology to retrieve sea surface currents and directional wave spectra. A procedure based on the dispersion relation, which was previously applied to process X-band marine radar data, is here exploited. The estimation performance of the radar sensor is first assessed by numerical tests in the case of synthetic sea wave fields with known characteristics in terms of wave direction and surface currents. Finally, the estimation procedure is assessed on real data collected at San Vincenzo quay in the port area of Naples, Italy. The achieved results are encouraging and highlight that 24 GHz FMCW MIMO radar is a viable technology for sea wave monitoring.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Enhancing Forward-Looking Image Resolution: Combining Low-Rank and
           Sparsity Priors

    • Free pre-print version: Loading...

      Authors: Junkui Tang;Zheng Liu;Lei Ran;Rong Xie;Jikai Qin;
      Pages: 1 - 12
      Abstract: Compressed sensing (CS)-based imaging technology has attracted a lot of interest because it can enhance imaging resolution. Targets of interest for forward-looking imaging radar are typically few in comparison to the entire imaging region. This sparsity allows for the natural application of CS to the reconstruction of high-resolution forward-looking images. However, conventional CS-based imaging methods can only perform well when the signal-to-noise ratio (SNR) is high. Strong noise in radar imaging prevents the CS-based methods from producing excellent imaging results. Inspired by the low-rank property of the received radar target echo and the sparsity of the forward-looking image targets, we present a combined low-rank and sparse prior restricted model for forward-looking imaging with a multichannel array radar to overcome strong noise. To solve the low-rank joint sparse double prior constraint optimization problem, an augmented Lagrange multiplier (ALM) reconstruction method under the framework of the alternating direction multiplier method (ADMM) is proposed. Finally, the results of simulation and real measurement data indicate that our presented method is fairly effective at enhancing the azimuth resolution and robustness of forward-looking radar imaging in comparison to other current methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Novel Method for Staggered SAR Imaging in an Elevation Multichannel
           System

    • Free pre-print version: Loading...

      Authors: He Huang;Penghui Huang;Yanyang Liu;Huaitao Fan;Yunkai Deng;Xingzhao Liu;Guisheng Liao;
      Pages: 1 - 19
      Abstract: Synthetic aperture radar (SAR) is an advanced remote sensing technique, capable of observing Earth’s surface independent of weather conditions and sunlight illumination. Restricted by the minimum antenna area, however, conventional spaceborne SAR systems cannot achieve high azimuth resolution in a wide swath. In addition, blind ranges are present as the constant pulse repetition interval (PRI) is used. To solve these problems, a PRI-staggered elevation multichannel SAR (EMC-SAR) system is employed in this article. By transmitting the continuously PRI-varied sequence, the blind ranges are located at different regions in different receive instants, effectively avoiding the loss of coverage in elevation. In this system, three issues are required to be addressed: 1) recovering the missed data located at blind ranges; 2) suppressing range ambiguous components; and 3) restoring the PRI-varied signal into a regular grid. To deal with these problems, we propose a novel SAR imaging method for a PRI-staggered EMC-SAR system. To be applied on-ground, assume downlinking of the individual elevation channels. First, the modified $varepsilon $ -insensitive loss tube regression with the L2 regularization method is applied to recover the missed data. Then, the range ambiguous components are suppressed by performing digital beamforming (DBF) based on the elevation multichannel technique, where the covariance matrix is constructed by using an iterative adaptive algorithm. After that, a generalized scaling transform is employed to restore the PRI-varied signal into a uniform sampled grid. Finally, a well-focused SAR image can be obtained by performing the conventional SAR imaging techniques. The effectiveness of the proposed method is validated by both simulated and real SAR data processing results.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A System Optimization Scheme for Bias Correction of Polarimetric
           Phased-Array Radar

    • Free pre-print version: Loading...

      Authors: Yaomin He;Tao Zhang;Huafeng He;Junjun Yin;Jian Yang;
      Pages: 1 - 17
      Abstract: With the change in the spatial angle, the cross-polarization isolation (CPI) of polarimetric phased-array radar (PPAR) changes as well, destroying the estimation of the target polarization scattering matrix (PSM). To correct the bias in PPAR, this article comprehensively designs the transmitting antenna, receiving antenna, and signal waveform and proposes a bias correction method based on system optimization. First, for the transmitting antenna of PPAR, the second-order cone program (SOCP) model is proposed to optimize the weighting coefficient. With the SOCP-based beamforming method, not only beam pattern in any spatial angle can be achieved, but also arbitrary polarization state can be precisely configured. Then, for the wideband receiving signal with a certain beamwidth, an angle estimation method based on eigenvalue decomposition is proposed in this article, which can effectively cure the challenges introduced by the beamwidth and signal bandwidth. Subsequently, for the signal waveform, the phase code is designed to measure all the elements of PSM in simultaneous transmission and simultaneous reception (STSR) mode, which could eliminate the biases of the moving speed and the second-order cross-polarization error. Finally, this article compares with other methods based on differential reflectivity, and experiments show that the factors such as the spatial angle, array structure, antenna beamwidth, signal bandwidth, motion speed, and signal-to-noise ratio (SNR) have the least influence on the present method in this article.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multi-UAV Collaborative Trajectory Optimization for Asynchronous 3-D
           Passive Multitarget Tracking

    • Free pre-print version: Loading...

      Authors: Jinhui Dai;Wenqiang Pu;Junkun Yan;Qingjiang Shi;Hongwei Liu;
      Pages: 1 - 16
      Abstract: This article considers the 3-D collaborative trajectory optimization (CTO) of multiple unmanned aerial vehicles to improve multitarget tracking performance with an asynchronous angle of arrival measurements. The predicted conditional Cramér–Rao lower bound is adopted as a performance measure to predict and subsequently control tracking error online. Then, the CTO problem is cast as a time-varying nonconvex problem subjected to constraints arising from dynamic and security (height, collision, and obstacle/target/threat avoidance). Finally, a comprehensive solution method (CSM) is presented to tackle the resulting problem, according to its unique structures. Specifically, if all security constraints are inactive, the CTO can be simplified as a nonconvex problem with convex dynamic constraints, which can be solved by the nonmonotone spectral projected gradient (NSPG) method. Oppositely, an alternating direction penalty method (ADPM) is presented to solve the CTO problem with some positive security constraints. The ADPM introduces auxiliary vectors to decouple the complex constraints and separates the CTO into several subproblems and tackles them alternately, while locally adjusting the penalty factor at each iteration. We show the subproblem w.r.t. the position vector is nonconvex but with convex constraints, which can be efficiently solved by the NSPG method. The subproblems w.r.t. the auxiliary vectors are separable and have closed-form solutions. Simulation results demonstrate that the CSM outperforms the unoptimized method in terms of tracking performance. Besides, the CSM achieves the near-optimal performance provided by the genetic algorithm with much lower computational complexity.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Inversion of Wave Parameters From Time-Doppler Spectrum Using Shore-Based
           Coherent S-Band Radar

    • Free pre-print version: Loading...

      Authors: Chen Zhao;Xiao Wang;Zezong Chen;Sitao Wu;Yichen Zeng;
      Pages: 1 - 11
      Abstract: For coherent microwave radar, the Bragg scattering from the broken-short waves generated after wave breaking usually introduces extra low-frequency components in the estimated wave height spectrum and, thus, leads to inaccurate retrievals of wave parameters, especially the overestimation of wave period. In order to eliminate the impacts of wave breaking, some methods based on the removal of “group line” in the spatial–temporal domain are proposed to estimate wave parameters. However, these methods are not suitable for the case that the spatial–temporal data are not available. To address this problem, a method is proposed to invert wave parameters only from the time-Doppler spectrum. Temporal velocity series are derived from the time-Doppler spectrum in which breaking components are removed. Then, the wave height spectrum from which wave parameters can be obtained is estimated from the velocity series by the direct transform relationship based on the linear wave theory. Without spatial–temporal data, the “group line” can be removed using the proposed method, and the method is validated by simulation. In addition, an approximately 11-day dataset collected with a shore-based coherent S-band radar deployed along the coast of Zhejiang province in China is reanalyzed and used to retrieve significant wave height and mean wave period. Compared with the buoy-measured data, the significant wave heights and mean wave periods retrieved by the proposed method have the root-mean-square differences (RMSDs) of 0.25 m and 0.60 s, respectively, and also have the correlation coefficients (CCs) of 0.94 and 0.81, respectively. The results indicate that the proposed method can invert wave parameters from the time-Doppler spectrum with a reasonable performance.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Marine Radar Small Target Classification Based on Block-Whitened
           Time–Frequency Spectrogram and Pre-Trained CNN

    • Free pre-print version: Loading...

      Authors: Shuwen Xu;Hongtao Ru;Dongchen Li;Penglang Shui;Jian Xue;
      Pages: 1 - 11
      Abstract: This article presents a classification method to classify different marine floating small targets, which can realize effective classification of different targets in strong clutter background. The design of proposed classification method is primarily based on block-whitened time–frequency spectrogram and pre-trained convolution neural network (CNN). Block-whitening clutter suppression is used to process target echoes. By converting a strong clutter background to an approximately noisy background, the effect of strong clutter on classification is reduced. Then, the time–frequency spectrogram of targets is extracted from the block-whitened target echoes, which converts a signal in time domain into a time–frequency spectrogram with more information. In addition, the block-whitened time–frequency spectrograms are input to a pre-trained CNN for feature extraction and classification training. By exploiting pre-training procedure, the proposed method can effectively classify different marine floating small targets and solve the problem of limited target samples in practical applications. Finally, a dataset of three kinds of measured maritime radar targets is constructed to verify the effectiveness of proposed method. Experimental results show that compared with competitors, the pre-trained CNN with block-whitened time–frequency spectrograms can achieve higher performance on the measured dataset.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Trihedral Corner Reflector for Radar Altimeter Calibration

    • Free pre-print version: Loading...

      Authors: Ferran Gibert;Adrià Gómez-Olivé;Albert Garcia-Mondéjar;Richard Francis;Sergi Hernández;Adrián Flores de la Cruz;Ester Vendrell;Mònica Roca i Aparici;
      Pages: 1 - 8
      Abstract: A trihedral corner reflector has been used to evaluate the capability of passive reflectors to calibrate radar altimeters, such as the Poseidon-4 altimeter on board Sentinel- 6A. The reflector location, placed on the top of a mountain ridge and about 4-km off-nadir of the Sentinel- 6A subsatellite track, allows capturing echoed signals with a signal-to-clutter ratio (SCR) around 40 dB when processing the received data with a fully-focused synthetic aperture radar (FF-SAR) algorithm. Results obtained show a range bias of 33.9 with 8.5 mm of standard deviation and datation bias of $- 2.3,,mu text{s}$ with a standard deviation of $1.8~mu text{s}$ for the measurement campaign between September 2021 and April 2022. Such values are comparable to what is currently achieved by means of active transponders, and therefore, it is demonstrated that passive reflectors may be of interest to support radar altimeter regular calibration.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Deep-Learning-Based Flying Animals Migration Prediction With Weather Radar
           Network

    • Free pre-print version: Loading...

      Authors: Huafeng Mao;Cheng Hu;Rui Wang;Kai Cui;Shuaihang Wang;Xiao Kou;Dongli Wu;
      Pages: 1 - 13
      Abstract: Monitoring and forecasting aerial animal migration benefit biological conservation, aviation safety, and agricultural production. Due to the lack of large-scale observation data and quantitative knowledge of aerial animal migration mechanisms, it is difficult to build a numerical simulation system for migration prediction. However, the extensive deployment of weather radars makes it possible to obtain large-scale aerial migration information. Meanwhile, artificial intelligence technologies provide new insights into the modeling of complex system. In this article, we develop a deep-learning model to predict aerial migration from the perspective of spatio-temporal evolution. Specifically, an undirected graph is applied to describe the geographic structure of the weather radar network, and then graph convolution and gated recurrent unit (GRU) are combined to extract spatio-temporal features of migration information. In addition, a multi-head self-attention mechanism is applied to enhance long-term dependence. Experiments are conducted to validate the effectiveness of the proposed model on the data from the Chinese weather radar network. The results show that our model can achieve state-of-the-art performance among the competing methods. Moreover, improvements from graph convolution and multi-head self-attention are also analyzed. In future applications, more weather radar data will be collected to enrich the dataset and build an aerial migration monitoring and prediction system.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Novel Approach for Radar Passive Jamming Based on Multiphase Coding
           Rapid Modulation

    • Free pre-print version: Loading...

      Authors: Heng Xu;Yinghui Quan;Xiaoyang Zhou;Hui Chen;Tie Jun Cui;
      Pages: 1 - 14
      Abstract: Electromagnetic controlled surface (ECS) which can regulate the amplitude and phase of electromagnetic wave reflection characteristics has attracted extensive attention in recent years because of its radar stealth effect by reducing the radar cross Section (RCS). However, radar detection cannot be avoided only by stealth in the energy field gradually. It is a novel pulse Doppler radar jamming mode proposed in this article to use the self-correlation of radar-transmitted waveforms to result in signal processing and radar detection failure through the rapid time-domain variation of phase coding ECS. In principle, phase modulation alters the intrapulse characteristics of the original signal and fundamentally interferes with the processing of the radar signal. In this article, random and periodic phase coding sequences are proposed according to different forms of radar jamming. And through the derivation formula and simulation experiment, the jamming effect under multiphase modulation is verified. Among them, the target energy is dispersed to the surroundings to form a wide envelope through random coding modulation, which leads to noise barrage jamming. While we use periodic sequence coding modulation, the target is shifted in the range–Doppler domain, causing misplaced coherence on the radar echo and deceptive jamming to the radar. Moreover, the technology can also be widely used in synthetic aperture radar (SAR) imaging, microwave measurement, and other remote sensing fields. To further confirm the accuracy and efficiency of the proposed method, multiparameter contrast experiments and parameter sensitivity analyses are conducted.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Sparse Aperture Autofocusing and Imaging Based on Fast Sparse Bayesian
           Learning From Gapped Data

    • Free pre-print version: Loading...

      Authors: Yuanyuan Wang;Fengzhou Dai;Qian Liu;Ling Hong;Xiaofei Lu;
      Pages: 1 - 16
      Abstract: Sparse aperture (SA) autofocusing and imaging is a hot research problem in the signal processing field and has been widely used. Under SA, the absence of echoes destroys the coherence between the pulses, which then affects the autofocusing accuracy of the imaging, leading to defocus of the image. In this article, a novel SA autofocusing and imaging algorithm based on sparse Bayesian learning (SBL) is proposed, which uses a fast SBL algorithm to achieve SA high-resolution imaging and the minimum Tsallis entropy algorithm to realize autofocusing. As is known to all, SBL has strong robustness and high precision. Unfortunately, the direct calculation of the inversion and multiplication operations involved in each iteration of SBL results in significant computational costs. In the proposed fast SBL algorithm, the matrix required to be inverted has a special structure. The inverse matrix can then be represented by Gohberg–Semencul (G–S) factorization. Also, almost all operations except for G–S factorization during each iteration can be completed by fast Fourier transform (FFT) or inverse FFT (IFFT), which greatly reduces the amount of computation by several orders of magnitude. In each SBL iteration, the minimum Tsallis entropy algorithm is used for estimating the phase error, which has better noise sensitivity and obtains the images with the best focused degree. Finally, the effectiveness and high efficiency of the proposed fast algorithm are verified by experimental results obtained by simulation and measured data.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • LDA-MIG Detectors for Maritime Targets in Nonhomogeneous Sea Clutter

    • Free pre-print version: Loading...

      Authors: Xiaoqiang Hua;Linyu Peng;Weijian Liu;Yongqiang Cheng;Hongqiang Wang;Huafei Sun;Zhenghua Wang;
      Pages: 1 - 15
      Abstract: This article deals with the problem of detecting maritime targets embedded in nonhomogeneous sea clutter, where the limited number of secondary data is available due to the heterogeneity of sea clutter. A class of linear discriminant analysis (LDA)-based matrix information geometry (MIG) detectors is proposed in the supervised scenario. As customary, Hermitian positive-definite (HPD) matrices are used to model the observational sample data, and the clutter covariance matrix of the received dataset is estimated as the geometric mean of the secondary HPD matrices. Given a set of training HPD matrices with class labels, which are elements of a higher dimensional HPD matrix manifold, the LDA manifold projection learns a mapping from the higher dimensional HPD matrix manifold to a lower dimensional one subject to maximum discrimination. In this study, the LDA manifold projection, with the cost function maximizing between-class distance while minimizing within-class distance, is formulated as an optimization problem in the Stiefel manifold. Four robust LDA-MIG detectors corresponding to different geometric measures are proposed. Numerical results based on both simulated radar clutter with interferences and real IPIX radar data show the advantage of the proposed LDA-MIG detectors against their counterparts without using LDA and the state-of-the-art maritime target detection methods in nonhomogeneous sea clutter.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • NLOS Positioning for Building Layout and Target Based on Association and
           Hypothesis Method

    • Free pre-print version: Loading...

      Authors: Peilun Wu;Jiahui Chen;Shisheng Guo;Guolong Cui;Lingjiang Kong;Xiaobo Yang;
      Pages: 1 - 13
      Abstract: Localization of non-line-of-sight (NLOS) targets in the complex urban environment has attracted significant attention in recent years. However, the requirement for precise prior information about the environment is idealistic. It is challenging to know the environmental information in the blind area of vision in advance of practical applications. This article proposes a joint estimation algorithm for building layout and target position in the L-shaped scene without any prior information. Specifically, a round-trip multipath propagation model is first developed for the cases of diffraction and multiple reflections. Then, the received echo signal is preprocessed with moving target identification (MTI), back-projection (BP) imaging, and image segmentation. In addition, the target points, which are screened by the geometric association, are further matched and estimated by the multipath ghost’s hypothesis method, thus realizing the joint perceptual estimation of the building layout and the target position. Finally, electromagnetic (EM) simulations and experimental measurements are used to validate the effectiveness of the proposed algorithm.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Track-to-Track Association Based on Maximum Likelihood Estimation for
           T/R-R Composite Compact HFSWR

    • Free pre-print version: Loading...

      Authors: Weifeng Sun;Xiaotong Li;Zhenzhen Pang;Yonggang Ji;Yongshou Dai;Weimin Huang;
      Pages: 1 - 12
      Abstract: Due to its low transmit power and reduced aperture size of a receiving antenna array, compact high-frequency surface wave radar (HFSWR) suffers from low detection probability, low positioning accuracy, and high false alarm rate. In a multitarget tracking scenario, similar kinematic parameters of adjacent targets raise challenges to the track-to-track association procedure. Taking the measurement uncertainty of compact HFSWR into consideration, a track-to-track association method based on maximum likelihood estimation (MLE) for T/R-R composite compact HFSWR is proposed. First, a multitarget tracking algorithm is applied to plot data sequences acquired by both T/R monostatic and T-R bistatic radars to produce two track sets. Then, the measurement errors of range, azimuth, and Doppler velocity are calculated using the obtained radar track and corresponding automatic identification system (AIS) track data, and a Gaussian distribution model is derived through probability distribution fitting. Subsequently, likelihood functions are established using the obtained Gaussian distribution model to calculate the association cost of tracks for T/R monostatic and T-R bistatic radars, and a cost matrix is obtained. Finally, the Jonker–Volgenant–Castanon (JVC) assignment algorithm is applied to the cost matrix to determine associated track-to-track pairs. Track-to-track association experiments using both simulated and field data were conducted, and the association performance of the proposed method is compared with that of Mahalanobis distance-based nearest neighbor (NN) method. Experimental results demonstrate that the proposed method can effectively resolve association ambiguity and achieve correct track-to-track association in track crossing and adjacent multitarget scenarios.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Attenuation Correction in Weather Radars for Snow

    • Free pre-print version: Loading...

      Authors: Shashank S. Joshil;V. Chandrasekar;
      Pages: 1 - 14
      Abstract: Weather radars play a prominent role in remote sensing of the atmosphere. Various fields, such as meteorology and hydrology, rely on accurate weather radar data as input for their models. Different hydrometeors present during a weather event influence the amount of attenuation encountered by the radar signal. Attenuation correction for dual-polarization weather radar data is necessary to improve the radar products and get accurate measurements. Most of the existing attenuation correction research is associated with rain hydrometeors. Currently, research that addresses the attenuation correction of snow in weather radars is limited. Although it is known that attenuation of radar signals when it encounters rain is much greater than that for snow, attenuation for all hydrometeors needs to be addressed for accurate radar estimates. In this research work, the attenuation of different hydrometeors is studied using signal simulations. Various factors which influence attenuation, such as the elevation angle and particle size distribution, are considered, and the results are presented. An attenuation correction algorithm that uses the hydrometeor classification and specific differential phase products from the DROPS2.0 algorithm is introduced. Signal simulations are employed to obtain the relationship between specific attenuation and specific differential phases for different hydrometeors used in the proposed algorithm. The attenuation correction method is applied to X-band and Ku-band radar data. Path-integrated attenuation of about 8 dB was observed in the snow case discussed from Ku-band radar data. The method proposed for attenuation shows promising results at both frequency bands.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Joint Design of Transmit Sequence and Receive Filter Based on Riemannian
           Manifold of Gaussian Mixture Distribution for MIMO Radar

    • Free pre-print version: Loading...

      Authors: Xixi Chen;Hao Wu;Yongqiang Cheng;Weike Feng;Yifan Guo;
      Pages: 1 - 13
      Abstract: To improve target detection performance in non-Gaussian backgrounds, the joint design of transmit sequence and receive filter for multiple-input–multiple-output (MIMO) radar is studied. By approximating the probability density function of observed non-Gaussian data with the Gaussian mixture model (GMM), a Riemannian manifold of Gaussian mixture distribution is developed to depict the complicated background first. Then, maximizing the geometric distance on manifolds, which is converted by maximizing the discrimination between the target and clutter, is proposed as the criterion for the joint design of transmit sequence and receive filter. Thereby, under the constant-modulus constraint, the joint design problem can be transformed into an optimization problem. However, the proposed optimization problem is nonconvex and constrained. To solve this problem, a Riemannian optimization framework is provided. By taking the advantage of the underlying geometric and algebraic structure of the constraint space, the original constrained optimization problem in Euclidean space can be transformed into the unconstraint optimization problem over Riemannian product manifolds. Moreover, to obtain the global optimal solution, the Riemannian gradient of the geometric distance cost is derived for the conjugate gradient algorithm. Experiments demonstrate that the proposed method shows advantages in detection performance compared with competitive methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Variable Length Sequential Iterable Convolutional Recurrent Network for
           UWB-IR Vehicle Target Recognition

    • Free pre-print version: Loading...

      Authors: Di Chen;Gang Xiong;Lizhe Wang;Wenxian Yu;
      Pages: 1 - 11
      Abstract: A variable length sequential iterable convolutional recurrent network (VS-ICRN) is proposed in this article, aiming at improving the vehicle target recognition ability for the ultrawideband impulse radar (UWB-IR). First, the array imaging technology is introduced into the UWB-IR, and thus, a range-angle imaging method for the array UWB-IR is put forward, to simulate the array UWB-IR vehicles image under different observation conditions. Second, in order to make full use of both the deep features in the single image and the deep associated features between the sequence images, a VS-ICRN model is proposed, which includes three submodules: the image feature extraction based on the iterable convolution, the variable length sequential image associated feature extraction, and the target classification. Finally, the experiment on the simulation dataset and the Moving and Stationary Target Acquisition and Recognition (MSTAR) is carried out to validate the effectiveness of the proposed method. The experimental results on the simulation dataset show that when SNR $= -10$ dB, the proposed method is superior in the recognition rate to the GoogLeNet and AlexNet methods with 16% and 19%, respectively. Meanwhile, the proposed VS-ICRN method only needs 1.38% parameters quantity to achieve a comparable recognition rate as GoogLeNet on the MSTAR dataset.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Generalized Geodesic Distance-Based Approach for Analysis of SAR
           Observations Across Polarimetric Modes

    • Free pre-print version: Loading...

      Authors: Debanshu Ratha;Andrea Marinoni;Torbjørn Eltoft;
      Pages: 1 - 16
      Abstract: Present and future sensors are diversifying from traditional quad polarimetric mode of synthetic aperture radar acquisition. Thus, an approach that is interpretative in nature and applicable across polarimetric modes is required. In this context, the geodesic distance (GD)-based approach within the polarimetric synthetic aperture radar (PolSAR) literature is seen as an eigenvalue-decomposition free approach to interpret and analyze quad PolSAR data. This approach is highly adaptive toward applications due to its ability to compare the SAR observation with a known scatterer/model, or with another SAR observation in general providing a means for direct interpretation. In this work, we show that the GD (originally defined for the quad polarization mode) is generalizable across any arbitrary SAR polarimetric mode while retaining its simple form for ready computation. We show that the GD-based approach provides level ground for comparison of different polarimetric modes given a fixed application. We demonstrate it using change detection as the chosen application. In addition, we show how the behavior of the three roll-invariant GD-based parameters change under different polarimetric modes (e.g., quad, dual, and compact polarization modes). We also discuss how the GD-based approach can also be adapted to ground range detected (GRD) product data, which is presently available from Sentinel-1 and widely used in many applications. However, in this case, we show that only one of the three GD-derived parameters can be defined. We believe this work will make the GD-based approach important for present and future PolSAR applications cutting across sensors and its available polarimetric modes.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Waveform Design for Watermark Framework Based DFRC System With Application
           on Joint SAR Imaging and Communication

    • Free pre-print version: Loading...

      Authors: Jing Yang;Youshan Tan;Xianxiang Yu;Guolong Cui;Di Zhang;
      Pages: 1 - 14
      Abstract: In this article, the watermarking framework for electromagnetic systems is established with nonblind, semiblind, and blind watermarking demodulation processes mitigated to applications like radar detection, synchronization, integrated dual function, etc. Desired for similar advantages and trade-offs of watermarking technology, the dual function radar and communication (DFRC) system is specifically concerned for covert communication and low possibility of interception (LPI) radar sensing. In this respect, we propose a novel DFRC waveform design method where peak sidelobe level (PSL) of autocorrelation function (ACF) is considered as the figure of merit with information embedded via discrete Fourier transform (DFT) watermarking strategy. Meanwhile, the peak-to-average ratio (PAR) and energy constraints are forced to ensure compatibility with current hardware technique. To handle the resulting NP-hard design problems, the proximal method of multipliers (PMM) is employed with the overall computational burden linear with the amount of information per pulse and quadratic with respect to the code length. Finally, numerical and experimental results are provided to evaluate the effectiveness of the proposed DFRC waveform design scheme with application in joint synthetic aperture radar (SAR) imaging and communication.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Parameter-Free Enhanced SS&E Algorithm Based on Deep Learning for
           Suppressing Azimuth Ambiguities

    • Free pre-print version: Loading...

      Authors: Yuxi Suo;Kun Fu;Youming Wu;Wenhui Diao;Wei Dai;Xian Sun;
      Pages: 1 - 16
      Abstract: Aliasing artifacts introduced by azimuth ambiguity seriously impact the interpretation of synthetic aperture radar images. To achieve parameter-free and fast azimuth ambiguity suppression, a novel deep learning model is designed to estimate the ambiguous signal intensity to total signal intensity ratio in the range-Doppler domain. This model does not depend on processing parameters and can be applied in any acquisition mode. The mean shift algorithm is applied to select less ambiguous subspectra according to the estimation result. The selected subspectra are restored to a full spectrum with an energy concentrated extrapolation method to preserve the resolution. The enhanced spectral selection and extrapolation algorithm overcomes the dependence on processing parameters, and experiments based on TerraSAR-X and Radarsat-2 images indicate that the proposed algorithm suppresses the azimuth ambiguity significantly.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • SAR Imaging in Frequency Scan Mode: System Optimization and Potentials for
           Data Volume Reduction

    • Free pre-print version: Loading...

      Authors: Nicola Gollin;Rolf Scheiber;Michele Martone;Paola Rizzoli;Gerhard Krieger;
      Pages: 1 - 20
      Abstract: Frequency scanning (FScan) is an innovative acquisition mode for synthetic aperture radar (SAR) systems. The method is based on the frequency-dependent beam pointing capabilities of phased array antennas, artificially increased via the combined use of true time delays and phase shifters within the array antenna. By this, typical limitations of conventional SAR systems in terms of achievable swath width and azimuth resolution can be mitigated, and so a wide swath can be imaged maintaining a fine azimuthal resolution. In the first part of the article, we introduce the theoretical concept, which is necessary to evaluate the reduced echo window length (EWL) with respect to equivalent stripmap data and the implications for the transmit pulse characterization. An FScan sensor flying in a TerraSAR-X-like orbit is shown to be capable of imaging an 80-km wide swath with 1-m azimuth resolution. The resulting time–frequency properties of the recorded raw data make the traditional SAR data compression algorithms such as block-adaptive quantization (BAQ) highly inefficient in this case. Therefore, the second part of the article investigates dedicated quantization methods for efficient data volume reduction in FScan systems. Different solutions are investigated and evaluated through simulations. Various transformations of the raw data have been exploited to optimize the encoding process, including deramping, fast Fourier transform (FFT), and blockwise approaches. Compared with standard BAQ in the time domain, the suggested data compression methods significantly improve the resulting signal-to-quantization noise ratio, allowing for the reduction in the overall data volume by about 60% for the considered system scenario, while maintaining robustness in the presence of inhomogeneous scene characteristics at the cost of a modest complexity increase for its on-board implementation.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Moving Targets Detection for Video SAR Surveillance Using Multilevel
           Attention Network Based on Shallow Feature Module

    • Free pre-print version: Loading...

      Authors: He Yan;Xing Xu;Guodong Jin;Qianru Hou;Zhe Geng;Ling Wang;Jindong Zhang;Daiyin Zhu;
      Pages: 1 - 18
      Abstract: In this article, a novel method for the moving target detection through multilevel spatial and channelwise attention network based on shallow feature channel module (MSCA-SFCM) is presented, and the circular spotlight (CSL) video synthetic aperture radar ground moving target indication (Video-SAR-GMTI) mode of the Nanjing University of Aeronautics and Astronautics miniature SAR (NUAA MiniSAR) system is introduced. However, due to the lack of moving target samples, MSCA-SFCM cannot be directly applied to the CSL Video-SAR-GMTI mode in the real system. To this end, this article proposes a training sample library construction scheme for moving targets of high verisimilitude. In this scheme, based on the radar system parameters, after the traversal of moving target parameters and SAR imaging, the scattering line characteristic of all possible moving targets under the current system parameters is simulated and then used for MSCA-SFCM network training. Afterward, the properly trained network can be used for moving target detection in real radar data. The effectiveness of the proposed method is verified by the NUAA MiniSAR system.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Resolution Enhancement for Forwarding Looking Multi-Channel SAR Imagery
           With Exploiting Space–Time Sparsity

    • Free pre-print version: Loading...

      Authors: Jingyue Lu;Lei Zhang;Shaopeng Wei;Yachao Li;
      Pages: 1 - 17
      Abstract: Forward-looking multi-channel synthetic aperture radar (FLMC-SAR) is of the capability to achieve unambiguous 2-D images in the forward-looking slight direction. FLMC-SAR imagery usually suffers from relatively low spatial resolution as only limited Doppler diversity can be generated from the synthetic aperture. In this article, a sparsity-driven resolution enhancement algorithm is proposed to improve the resolution FLMC-SAR image of the forward-looking area. Different from conventional beamforming processing to resolve the FLMC-SAR left–right ambiguity, a Bayesian sparsity reconstruction optimization is developed for jointly ambiguity resolving and resolution enhancement in the azimuth angle image domain. The spatial structure of the target in the preliminary image domain is used as the signal sparsity with prior information to solve the constrained optimization problem for FLMC-SAR image resolution enhancement. A local least square estimator of the prior noise and signal statistics in the FLMC-SAR nonisotropic image is established in terms of determining the sparsity weight parameter. Extensive simulation and real FLMC-SAR data experiments confirm that the proposed algorithm is capable of achieving the unambiguous and resolution-enhanced FLMC-SAR image.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Ground Moving Target Detection With Nonuniform Subpulse Coding in SAR
           System

    • Free pre-print version: Loading...

      Authors: Xiongpeng He;Yue Yu;Yifan Guo;Guisheng Liao;Shengqi Zhu;Jingwei Xu;Tong Gu;
      Pages: 1 - 18
      Abstract: For the high-resolution and wide-swath (HRWS) synthetic aperture radar (SAR) system, the increasing imaging width results in a serious range ambiguity problem, which affects the performance of ground moving target indication (GMTI). In this article, a novel nonuniform subpulse coding (NSPC) scheme is proposed. It is characterized by resorting to range-frequency band resources and detailed coding design for each subpulse, enabling the beam auto-scanning in elevation. Also, the bandpass filtering and digital beamforming (DBF) technology with improved data reconstruction are utilized to realize the separation of subpulses and suppress range ambiguity. The NSPC technique exchanges the signal bandwidth for increasing swath without range ambiguity, and the coded subpulses can be directed to the prescribed regions, while skipping the invalid areas where the echoes are blocked. After that, through the robust principal component analysis (RPCA) method, the moving target detection is performed for each separated region without residual range-ambiguous interference. The proposed approach has been theoretically deduced in detail and the simulation experiments demonstrate its effectiveness.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • What Catch Your Attention in SAR Images: Saliency Detection Based on
           Soft-Superpixel Lacunarity Cue

    • Free pre-print version: Loading...

      Authors: Fei Ma;Xuejiao Sun;Fan Zhang;Yongsheng Zhou;Heng-Chao Li;
      Pages: 1 - 17
      Abstract: In existing superpixel-wise saliency detection algorithms, superpixel generation often is an isolated preprocessing step. The performance of saliency maps is determined by the accuracy of superpixels to a certain extent. However, it is still a challenge to develop a stable superpixel generation method. In this article, we attempt to incorporate the superpixel generation and saliency calculation steps into an end-to-end trainable deep network. First, we employ a recently proposed differentiable superpixel generation method to over-segment the synthetic aperture radar (SAR) images, which outputs the possibility that the pixels assigned to neighbor superpixels (soft superpixel). In saliency calculation part, as one of our main contributions, we propose a differentiable and computationally simple saliency model, i.e., lacunarity cue. It is inspired by the fact that generally the backscattering intensity of regions of interest (ROIs) in SAR images irregularly fluctuates, while the areas with consistent pixels are often ignored as the clusters. We improve the pixelwise box differential dimension algorithm to measure the irregularity of scattering points in a superpixel. The superpixel generation and saliency calculation can be implemented under a unified deep network. Hence, the shapes of the superpixels can be iteratively adjusted according to the saliency maps until the ROIs are correctly detected. Experiments on real SAR images with different sizes and scenes show that the saliency maps can effectively highlight the target areas, thus outperforming the state-of-the-art saliency detection models.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Monostatic-Equivalent Algorithm via Taylor Expansion for BiSAR Ship Target
           Imaging

    • Free pre-print version: Loading...

      Authors: Guangzhao Qian;Yong Wang;
      Pages: 1 - 19
      Abstract: It is well known that the bistatic synthetic aperture radar (BiSAR) can obtain high-resolution imaging in the forward direction of radar, which is not possible with the conventional monostatic SAR systems. However, most studies on BiSAR imaging mainly focus on stationary targets, and the research on moving targets and ship targets is relatively scarce. In this article, a monostatic-equivalent (ME) algorithm via Taylor expansion for BiSAR imaging of ship target is proposed. The monostatic equivalence method based on Taylor expansion is adopted to approximate BiSAR to monostatic model; besides, the method based on the midpoint of baseline is also analyzed for comparison. Then, the hybrid SAR-ISAR technology is utilized for ship target imaging under the ME model. Finally, the experimental results are given to substantiate the effectiveness and robustness of the proposed algorithm.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Unpaired Speckle Extraction for SAR Despeckling

    • Free pre-print version: Loading...

      Authors: Huangxing Lin;Yihong Zhuang;Yue Huang;Xinghao Ding;
      Pages: 1 - 14
      Abstract: Speckle suppression is a critical step in synthetic aperture radar (SAR) imaging. Since speckle-free SAR images are inaccessible, supervised denoising methods are not suitable for this task. To exploit the strong capabilities of convolutional neural networks (CNNs), we propose Unpaired Speckle Extraction (SAR-USE), an unsupervised method for SAR despeckling. Our method utilizes unpaired SAR and clean optical images to extract “real” speckle for learning despeckling. First, a CNN that has never seen clean SAR images is employed to extract speckle from the SAR image. Then, the extracted speckle is multiplied with a random optical image to synthesize paired data for learning speckle removal. Through a Siamese network, speckle extraction and learning despeckling are performed alternately and promote each other. To make the extracted speckle more visually and statistically realistic, it is constrained by a noise correction module to be unit mean while maintaining spatial correlation. After convergence, the CNN is a good denoiser that can effectively extract speckle from SAR images. Experiments on synthetic datasets show that the denoising ability of the proposed method is as good as its supervised counterpart. More importantly, SAR-USE is very efficient for removing the spatially correlated speckle in real data that supervised learning methods cannot.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Self-Supervised SAR Image Registration With SAR-Superpoint and
           Transformation Aggregation

    • Free pre-print version: Loading...

      Authors: Bin Zou;Haolin Li;Lamei Zhang;
      Pages: 1 - 15
      Abstract: Owing to various factors, including severe speckle noise and orbit direction differences, performing multitemporal synthetic aperture radar (SAR) image registration with high accuracy and robustness may become difficult. Herein, an efficient self-supervised deep learning registration network for multitemporal SAR image registration, SAR-superpoint and transformation aggregation network (SSTA-Net), is proposed. The SSTA-Net consists of three parts: 1) the SAR-Superpoint detection network (SS-Net); 2) the transformation aggregation feature matching network (TA-Net); and 3) the unstable point removal module. Specifically, a pseudolabel generation method is adopted without additional annotations. It transfers the characteristics of real SAR data to synthetic data through a feature transition module, which can generate feature point labels for real SAR images for self-training SS-Net. Furthermore, a position–channel aggregation attention is proposed and embedded into the SS-Net to efficiently capture position and channel information and to increase the stability and accuracy of feature point identification. Finally, a unique transformation aggregation strategy is designed to improve the robustness of feature matching, and an unstable point removal module is adopted to eliminate the mismatched point pairs caused by orbit differences. Six sets of multitemporal SAR images were used to evaluate the registration performance of the SSTA-Net, and our model was also compared with the traditional and deep learning algorithms. The experimental results demonstrate that the SSTA-Net outperforms various state-of-the-art approaches for SAR image registration.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Uncertainty-Incorporated Ice and Open Water Detection on Dual-Polarized
           SAR Sea Ice Imagery

    • Free pre-print version: Loading...

      Authors: Xinwei Chen;K. Andrea Scott;Linlin Xu;Mingzhe Jiang;Yuan Fang;David A. Clausi;
      Pages: 1 - 13
      Abstract: Algorithms designed for ice–water classification of synthetic aperture radar (SAR) sea ice imagery produce only binary (ice and water) output typically using manually labeled samples for assessment. This is limiting because only a small subset of labeled samples are used, which, given the nonstationary nature of the ice and water classes, will likely not reflect the full scene. To address this, we implement a binary ice–water classification in a more informative manner considering the uncertainty associated with each pixel in the scene. To accomplish this, we have implemented a Bayesian convolutional neural network (CNN) with variational inference to produce both aleatoric (data-based) and epistemic (model-based) uncertainty. This valuable information provides feedback as to regions that have pixels more likely to be misclassified and provides improved scene interpretation. Testing was performed on a set of 21 RADARSAT-2 dual-polarization SAR scenes covering a region in the Beaufort Sea captured regularly from April to December. The model is validated by demonstrating: 1) a positive correlation between misclassification rate and model uncertainty and 2) a higher uncertainty during the melt and freeze-up transition periods, which are more challenging to classify. By incorporating the iterative region growing with semantics (IRGS) segmentation algorithm and an uncertainty value-based thresholding algorithm, the Bayesian CNN classification outputs are improved significantly via both numerical analysis and visual inspection.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Cooperative SAR-Communication System Using Continuous Phase Modulation
           Codes and Mismatched Filters

    • Free pre-print version: Loading...

      Authors: Maria-Elisavet Chatzitheodoridi;Abigael Taylor;Olivier Rabaste;Hélène Oriot;
      Pages: 1 - 14
      Abstract: The electromagnetic congestion due to the continuous growth of spectral demand has been skyrocketing for the past years. Joint radar-communication systems are, thus, attracting attention as they can alleviate spectrum occupancy by using the same bandwidth to perform both applications. In this context, a cooperative radar-communication system, which is a specific category that uses communication codes to both transmit information and perform radar missions, can be considered. Continuous phase-modulated (CPM) codes are considered in this article in order to generate high-resolution radar images from airborne radar. Since mitigating the sidelobe level energy is essential for good image quality, we resort here to optimized mismatched filters (MMFs). A fast algorithm is proposed to minimize the computational time of these filters. Simulated data are generated, as well as resynthesized synthetic aperture radar (SAR) images, and reconstructed from real chirp-based data using CPM codes and MMFs. Their performance is evaluated using different comparison tools and shows that the use of mismatched filtering and different messages embedded in the phase of each transmitted code provides enhanced image quality.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • CZT Algorithm for the Doppler Scale Signal Model of Multireceiver SAS
           Based on Shear Theorem

    • Free pre-print version: Loading...

      Authors: Mengbo Ma;Jinsong Tang;Haoran Wu;Peng Zhang;Mingqiang Ning;
      Pages: 1 - 12
      Abstract: In multireceiver synthetic aperture sonars (SASs), the traditional non-stop-hop-stop model normally ignores the intrapulse motion of sonar platform. However, such intrapulse motion brings the Doppler scale effect (DSE) to the echo signal, which will lead to the change in carrier frequency (CF) and frequency modulation (FM) rate of echo signal and may eventually result in image defocus. In this article, considering the intrapulse motion, the time-varying time delay and range history of multireceiver SAS were derived by introducing the range time into the range history, and then the Doppler scale signal model was established. After approximating the derived time-varying range history, the 2-D spectrum was derived by using the shear theorem. When compared with the traditional 2-D spectrum, an interesting phenomenon was found that the change in FM rate has no effect on the 2-D spectrum. Actually, only two additional terms, i.e., range cell migration (ARCM) and azimuth modulation (AAM), were introduced into the new 2-D spectrum due to the change in CF. Besides, the condition of neglecting the ARCM was given by further analyzing the 2-D spectrum, while the AAM is usually small and can be easily compensated together with the ARCM correction. Finally, a chirp-z transform (CZT) algorithm for the Doppler scale signal model was proposed to correct the ARCM, which can simultaneously correct the ARCM and compensate the AAM (although not required) in the 2-D frequency domain by phase multiplication. The effectiveness of the signal model and the proposed algorithm is verified by simulations and field data.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Semisupervised SAR Ship Detection Network via Scene Characteristic
           Learning

    • Free pre-print version: Loading...

      Authors: Yuang Du;Lan Du;Yuchen Guo;Yu Shi;
      Pages: 1 - 17
      Abstract: In recent years, target detection methods based on deep learning have achieved extensive development in synthetic aperture radar (SAR) ship detection. However, training such detectors requires target-level annotations of SAR images that are hard to be obtained in practice. To reduce the dependence of network training on expensive target-level annotations, we propose a novel semisupervised SAR ship detection network via scene characteristic learning. The proposed network focuses on utilizing the scene-level annotations of SAR images to improve the detection performance in the case of limited target-level annotations. Compared with the traditional fully supervised SAR ship detection network, the proposed network constructs a scene characteristic learning branch parallel with the detection branch. In the scene characteristic learning branch, a scene classification loss and a scene aggregation loss are designed to utilize the scene-level annotations. Under the constraint of these two losses, the feature extraction network can fully learn the scene characteristics of SAR images, thus enhancing its feature representation ability for ship targets and clutter. In addition, we propose a hierarchical test process from scene to target. After recognizing the scene types of input SAR images, we design different detection strategies for SAR images recognized as different scenes. The proposed test process can significantly reduce the inland and inshore false alarms, thus leading to higher detection performance. The experiments based on two measured SAR ship detection datasets demonstrate the effectiveness of the proposed method.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Spatio-Temporal Filtering Approach for Tomographic SAR Data

    • Free pre-print version: Loading...

      Authors: Karima Hadj-Rabah;Gilda Schirinzi;Ishak Daoud;Faiza Hocine;Aichouche Belhadj-Aissa;
      Pages: 1 - 13
      Abstract: Synthetic aperture radar tomography (TomoSAR) has recently received particular interest from the remote-sensing community, due to its ability to provide 3-D reconstructions of environments with complex structures. Unfortunately, different forms of decorrelations and processing errors affect the quality of the resulting 2-D/3-D images. One way to cope with the impact of these nuisances is to apply appropriate filtering to the interferometric data stack as a preprocessing step. The first obstacle to be dealt with, especially in urban areas, is to define a filter whose parameters have to be set in such a way as to improve smoothing capabilities while preserving edges. To this aim, the main objective of this article is twofold: 1) the application of a spatio-temporal contextual filter whose parameters depend on 3-D quality indicators of the multibaseline interferometric image stack and 2) evaluation of the denoising effect on the application of nonparametric spectral estimation and detection algorithms. For that, we consider several quantitative metrics to assess, on the one hand, the filtering performances, and, on the other hand, its impact on the reflectivity function recovered from conventional tomographic inversion and detection methods. Experimental results from a set of TerraSAR-X (TSX) images highlight the efficiency of the filtering process by improving the scatterers’ detection and height localization, of a man-made structure.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A ConvLSTM Neural Network Model for Spatiotemporal Prediction of Mining
           Area Surface Deformation Based on SBAS-InSAR Monitoring Data

    • Free pre-print version: Loading...

      Authors: Sheng Yao;Yi He;Lifeng Zhang;Wang Yang;Yi Chen;Qiang Sun;Zhan’ao Zhao;Shengpeng Cao;
      Pages: 1 - 22
      Abstract: The surface deformation caused by underground mining leads to damage to surface buildings and brings potential safety hazards and property losses. The demand for reliable prediction methods of surface deformation in mining areas is becoming increasingly significant. At present, most prediction methods are based on sampling points; however, these methods neglect to consider local and overall spatial features, and this oversight affects the spatial accuracy of prediction results. The data form of the prediction output is often discontinuous and not intuitive. In order to solve this problem, the spatiotemporal prediction method of surface deformation in mining areas is a very effective proposal. However, few scholars have proposed a solution based on this idea. In this study, a convolution long short-term memory (ConvLSTM) neural network for surface deformation spatiotemporal prediction based on time-series interferometric synthetic aperture radar (InSAR) is proposed to directly predict the overall spatial deformation of the surface in the mining area. First, based on Sentinel-1A images of the Jinchuan Mining Area, Jinchang, Gansu province, China, the time-series InSAR surface deformation data of the study area from January 2018 to October 2020 (81 scenes) are obtained using small baseline subset InSAR (SBAS-InSAR) technology. Because of the large value scale of surface deformation in the mining area, we propose a method to fit the maximum and minimum values of time-series deformation, respectively, and carry out piecewise numerical compression. Then, based on the ConvLSTM neural network layer, construct the spatiotemporal prediction model of time-series InSAR surface deformation. Support vector regression (SVR), multilayer perceptron (MLP) regression, and the gray model [GM (1,1)] are used as benchmark methods. The prediction results of our models are compared with the three benchmark methods. The comparison results show that the prediction effect of the -onvLSTM model with optimal input time steps is significantly better than the benchmark methods in the comprehensive performance of various evaluation metrics, especially for the metrics used to evaluate the error. This shows that the ConvLSTM model has relatively fine spatiotemporal prediction performance for time-series InSAR surface deformation. Based on the InSAR time-series deformation monitoring results of the Jinchuan Mining Area, we carry out the spatiotemporal prediction of surface deformation in the subsequent 50 time steps (600 days). The results agree with the natural deformation development law in band collection statistics and 3-D representation of deformation; moreover, the reliability of numerical spatial distribution is relatively high. This result can be used to intuitively evaluate the overall surface deformation of the mining area in the monitoring range, find potential hazards in time, and take measures to address these hazards quickly. At the same time, this research process also provides a new concept design for such problems.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Study of Hypersonic Vehicle-Borne SAR Imaging Under Plasma Sheath

    • Free pre-print version: Loading...

      Authors: Lihao Song;Bowen Bai;Xiaoping Li;Gezhao Niu;Yanming Liu;Yiding Mao;
      Pages: 1 - 13
      Abstract: Hypersonic vehicle-borne synthetic aperture radar (SAR) imaging has promising application prospects for remote sensing owing to the characteristics of fast response and flexible trajectory. However, a newly observed phenomenon that occurs during hypersonic flight, called the plasma sheath, seriously affects hypersonic vehicle-borne SAR imaging. Therefore, in this study, the effect of plasma sheath on hypersonic vehicle-borne SAR imaging is investigated systematically. First, a hypersonic vehicle-borne SAR signal model under plasma sheath is developed. Then, the phase shift and amplitude attenuation which are critical in the SAR signal model under plasma sheath are deduced. Moreover, the analytical formulas for the linear and quadratic phase errors are derived. Based on the developed signal model, point target response and an SAR image of an area under plasma sheath show that linear phase error will cause image shift, and quadratic phase error will lead to pulse broadening. In addition, the amplitude attenuation will make the point target response submerged in noise, and signal amplitude nonlinear distortion will result in an asymmetric distortion phenomenon. To further explore the effect of plasma sheath on hypersonic vehicle-borne SAR imaging, quantitative evaluation and analysis of SAR imaging degradation are conducted under different electron densities, carrier frequencies, and bandwidths. The evaluation results in this study will benefit the application of hypersonic vehicle-borne SAR imaging under plasma sheath, and the analytical formulas of linear and quadratic phase error provide the possibility to compensate for SAR imaging degradation under plasma sheath in the future.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multitemporal Speckle Reduction With Self-Supervised Deep Neural Networks

    • Free pre-print version: Loading...

      Authors: Inès Meraoumia;Emanuele Dalsasso;Loïc Denis;Rémy Abergel;Florence Tupin;
      Pages: 1 - 14
      Abstract: Speckle filtering is generally a prerequisite to the analysis of synthetic aperture radar (SAR) images. Tremendous progress has been achieved in the domain of single-image despeckling. Latest techniques rely on deep neural networks to restore the various structures and textures peculiar to SAR images. The availability of time series of SAR images offers the possibility of improving speckle filtering by combining different speckle realizations over the same area. The supervised training of deep neural networks requires ground-truth speckle-free images. Such images can only be obtained indirectly through some form of averaging, by spatial or temporal integration, and are imperfect. Given the potential of very high-quality restoration reachable by multitemporal speckle filtering, the limitations of ground-truth images need to be circumvented. We extend a recent self-supervised training strategy for single-look complex (SLC) SAR images, called MERLIN, to the case of multitemporal filtering. This requires modeling the sources of statistical dependencies in the spatial and temporal dimensions as well as between the real and imaginary components of the complex amplitudes. Quantitative analysis on datasets with simulated speckle indicates a clear improvement of speckle reduction when additional SAR images are included. Our method is then applied to stacks of TerraSAR-X images and shown to outperform competing multitemporal speckle filtering approaches. The code of the trained models and supplementary results are made freely available at https://gitlab.telecom-paris.fr/ring/multitemporal-merlin/.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Saliency-Based SAR Target Detection via Convolutional Sparse Feature
           Enhancement and Bayesian Inference

    • Free pre-print version: Loading...

      Authors: Jia-Cheng Ni;Ying Luo;Dan Wang;Jia Liang;Qun Zhang;
      Pages: 1 - 15
      Abstract: Traditional synthetic aperture radar (SAR) target detection methods use matched filtered SAR images as input, and the detection performance is restricted due to the high sidelobes and speckle noise of these images. Sparse SAR imaging methods developed in recent years provide the advantages of reducing sidelobes, noise, and clutter. The imaging results obtained with these methods could help improve the SAR target detection performance. In this article, to improve the target detection performance using sparse SAR images as input, we proposed a convolutional sparse feature enhancement method to meet the needs of Bayesian saliency detection. The proposed Bayesian saliency joint target detection method comprised the following three steps: first, to obtain sparse SAR images with continuous contours and fewer holes in the target area, we proposed a convolutional L1 sparse regularization method. Second, a regularization parameter optimization method was derived to quickly obtain optimal regularization parameters for saliency detection. Finally, target detection results were obtained through a superpixel-based Bayesian saliency joint detector. Extensive experiments verified that the proposed method could improve the SAR target detection accuracy in complex backgrounds.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • ST-Net: Scattering Topology Network for Aircraft Classification in
           High-Resolution SAR Images

    • Free pre-print version: Loading...

      Authors: Yuzhuo Kang;Zhirui Wang;Haoyu Zuo;Yidan Zhang;Zhujun Yang;Xian Sun;Kun Fu;
      Pages: 1 - 17
      Abstract: Aircraft classification in synthetic aperture radar (SAR) images plays a considerable role in global region management and surveillance. Recently, deep learning has been applied to solve the classification problem and made significant progress. Due to the imaging variability at different angles and component scattering discreteness in SAR images, previous works have had difficulty in achieving desirable classification results. To address these issues, we study the positional and semantic relationship between the scattering points and propose an innovative scattering topology network (ST-Net) in this article. First, considering the diversity of imaging results caused by different target attitude angles, we extract and transform the scattering cluster centers to update the information of various categories. It can guide the model to strengthen the discriminative features and mitigate the impact of imaging variability on classification performance. Second, a novel scattering topology module (STM) is introduced to model the spatial relationships and semantic information interaction of discrete scattering points. In this process, the topology relations and scattering characteristics are enhanced for further accurate classification. Third, context attention excitation (CAE) is designed to capture significant global and semantic information, which is conducive to suppressing background interference and reducing category confusion. In conclusion, the ST-Net is presented with the SAR imaging mechanism and the topology geometric representation of aircraft. We construct the SAR aircraft category dataset (SAR-ACD) and conduct extensive experiments on it to show the effectiveness of ST-Net, which illustrates that our method achieves superior classification performance.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Image-Based Baseline Correction Method for Spaceborne InSAR With External
           DEM

    • Free pre-print version: Loading...

      Authors: Qingyue Yang;Jili Wang;Yingjie Wang;Pingping Lu;Hongying Jia;Zhipeng Wu;Lu Li;Yinkai Zan;Robert Wang;
      Pages: 1 - 16
      Abstract: An accurate baseline of synthetic aperture radar (SAR) interferometry (InSAR) is an important parameter for the geodetic application of the InSAR data. Although some advanced SAR satellites have precise orbit determination, there are still many SAR satellites suffering from baseline inaccuracies, such as GF-3. In this article, an image-based estimator for baseline correction is proposed, which requires only the external digital elevation model (DEM) data. The idea of the method is to project the orbit error phase onto the phase components carrying the baseline error information, which is called orbit error phase bases in this article, and to correct the baseline according to the projection coefficients. Since the pure orbit error phase is unavailable, the residual phase of the interferogram is used to approximate the orbit error phase, and a series of processes are introduced to weaken the effect of this approximation. Both the simulated and real data from GF-3 SAR are used to validate the proposed method, and a comparison with the conventional nonlinear least-square and the latest proposed flat-Earth phase-based baseline refinement methods are made. The results indicated the superior accuracy and robustness of our method, especially in areas with higher relief and wider coverage.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Segmental Aperture Imaging Algorithm for Multirotor UAV-Borne MiniSAR

    • Free pre-print version: Loading...

      Authors: Yixiang Luomei;Feng Xu;
      Pages: 1 - 18
      Abstract: This article takes on the challenges of synthetic aperture radar (SAR) imaging for miniaturized SAR (MiniSAR) onboard a multirotor unmanned aerial vehicle (UAV). Several unique challenges are systematically analyzed, and a corresponding analytical phase error model is established, which accurately models the effects of both translational and rotational motions of UAVs. A segmental aperture imaging (SAI) algorithm, an autofocus algorithm based on strong scatterers, is proposed. It simply divides the platform trajectory into uneven segments, which are first independently focused with motion compensation and then stitched together to form a complete SAR image. Both the theoretical derivation of the signal model and the implementation of the imaging algorithm are presented. A simulation analysis with actual UAV trajectory and attitude data is conducted, which demonstrates the efficacy and performance of the proposed imaging algorithm. It shows that the ideal focusing effect can be achieved as evaluated by various metrics, and the proposed algorithm has superior performance compared to the subaperture phase gradient autofocus (PGA) and minimum entropy autofocus (MEA) methods. Finally, the multirotor-borne MiniSAR system FUSAR-Ku is used for experiments to verify the proposed algorithm. Experimental results show that the proposed algorithm can achieve the theoretical decimeter-resolution imaging performance as measured by various metrics.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • MAda-Net: Model-Adaptive Deep Learning Imaging for SAR Tomography

    • Free pre-print version: Loading...

      Authors: Yan Wang;Changhao Liu;Rui Zhu;Minkun Liu;Zegang Ding;Tao Zeng;
      Pages: 1 - 13
      Abstract: The compressive sensing (CS)-based tomographic SAR (TomoSAR) 3-D imaging method has the shortcoming of low efficiency, mainly represented in two aspects: first, the CS solver requires iterative calculation and hence is computationally expensive; second, the CS solver needs hyperparameters’ selection, which commonly requires cost-inefficient try-and-error attempts. Recently, the iterative CS solver is suggested to be replaced by a deep learning network for a tremendous processing speed improvement. However, the existing deep-learning-based TomoSAR imaging algorithms suffer from the problem of model inadaptability, i.e., being inadaptive to the observation model and the signal energy model and hence is low accuracy. This article proposes a new model-adaptive network (MAda-Net) to implement deep-learning-based TomoSAR 3-D imaging with a much improved processing accuracy. First, a new adaptive model-solving (AMS) module is introduced to solve the problem of the observation model inconsistency between the real spatially varying one and the approximately fixed one used by the network. Second, a new adaptive threshold-activation (ATC) module is introduced to solve the problem of signal energy model inconsistency between the real backscattered echo and the simulated echo for network training. The effectiveness of the proposed method has been verified by the computer simulations and the real unmanned aerial vehicle (UAV) SAR experiments.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Joint Clutter Suppression and Moving Target Indication in 2-D Azimuth
           Rotated Time Domain for Single-Channel Bistatic SAR

    • Free pre-print version: Loading...

      Authors: Junao Li;Zhongyu Li;Qing Yang;Yahui Wang;Jie Long;Junjie Wu;Wei Xia;Jianyu Yang;
      Pages: 1 - 16
      Abstract: Moving target indication (MTI) in bistatic synthetic aperture radar (BiSAR) is a promising task in both civilian and military fields. However, it suffers severe challenges under the influence of clutter. MTI in BiSAR mainly faces three challenges: 1) the range cell migration (RCM) of BiSAR is larger than that of monostatic synthetic aperture radar; 2) the clutter range-Doppler spectrum is severely extended; and 3) the clutter characteristic is closely related to geometrical configuration, with nonhomogeneity and nonstationarity. To solve these problems, based on single-channel BiSAR, a joint clutter suppression and MTI method is proposed. First, to ensure that the energy of an arbitrary target is concentrated in one range cell, RCM correction (RCMC) is completed by the first-order keystone transform (KT) and high-order RCMC. Then, an important step named increased dimension rotation (IDR) is applied, which mainly consists of two stages. One is increased dimension processing, which introduces a 2-D azimuth rotated time domain (ARTD). The other is rotation processing, which makes signals in one range cell rotated to 2-D ARTD. After that, the distribution characteristic between the moving target and stationary clutter in 2-D ARTD is analyzed. Next, an optimal filter for clutter suppression is designed according to the characteristics of signal distribution in 2-D ARTD. Furthermore, the corresponding inverse IDR processing and fractional Fourier transform (FrFT) are performed, and finally, MTI can be realized accurately. Generally, the proposed method has good robustness and low system complexity, and its effectiveness is proven by numerical simulations.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multipath in Automotive MIMO SAR Imaging

    • Free pre-print version: Loading...

      Authors: Marco Manzoni;Stefano Tebaldini;Andrea Virgilio Monti-Guarnieri;Claudio Maria Prati;
      Pages: 1 - 12
      Abstract: This article discusses the effect of multipath in automotive radar imaging under different sensor configurations. The study is motivated by the fact that radar technologies are becoming indispensable in the automotive scenario. Many applications such as collision avoidance systems, assisted parking, and driving assistance systems take advantage of radar technologies to accomplish their task. However, one of the main concerns about automotive radars is the possibility of detecting false targets due to multiple signal reflections. In this article, we show how different sensor layouts experience multipath differently. In particular, we demonstrate that with multiple-input multiple-output (MIMO) radars, what really matters is the physical positions of the transmitting and receiving antennas. The monostatic/bistatic equivalent configurations cannot be used to design a system and to simulate an acquisition in the presence of a multipath. We also demonstrate how vehicle-based MIMO-synthetic aperture radar (MIMO-SAR) imaging can generate a bi-dimensional aperture which significantly reduces multipath effects in the focused image, avoiding the detection of false targets. All the theoretical analyses are supported by several simulations where different sensor layouts are tested, and the capability of MIMO-SAR to reject multipath is validated.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • 3-D SAR Imaging via Perceptual Learning Framework With Adaptive Sparse
           Prior

    • Free pre-print version: Loading...

      Authors: Mou Wang;Shunjun Wei;Jun Shi;Xiaoling Zhang;Yongxin Guo;
      Pages: 1 - 16
      Abstract: Mathematically, 3-D synthetic aperture radar (SAR) imaging is a typical inverse problem, which, by nature, can be solved by applying the theory of sparse signal recovery. However, many reconstruction algorithms are constructed by exploring the inherent sparsity of imaging space, which may cause unsatisfactory estimations in weakly sparse cases. To address this issue, we propose a new perceptual learning framework, dubbed as PeFIST-Net, for 3-D SAR imaging, by unfolding the fast iterative shrinkage-thresholding algorithm (FISTA) and exploring the sparse prior offered by the convolutional neural network (CNN). We first introduce a pair of approximated sensing operators in lieu of the conventional sensing matrices, by which the computational efficiency is highly improved. Then, to improve the reconstruction accuracy in inherently nonsparse cases, a mirror-symmetric CNN structure is designed to explore an optimal sparse representation of roughly estimated SAR images. The network weights control the hyperparameters of FISTA by elaborated regularization functions, ensuring a well-behaved updating tendency. Unlike directly using pixelwise loss function in existing unfolded networks, we introduce the perceptual loss by defining loss term based on high-level features extracted from the pretrained VGG-16 model, which brings higher reconstruction quality in terms of visual perception. Finally, the methodology is validated on simulations and measured SAR experiments. The experimental results indicate that the proposed method can obtain well-focused SAR images from highly incomplete echoes while maintaining fast computational speed.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Information Reconstruction-Based Polarimetric Covariance Matrix for PolSAR
           Ship Detection

    • Free pre-print version: Loading...

      Authors: Tao Zhang;Sinong Quan;Wei Wang;Weiwei Guo;Zenghui Zhang;Wenxian Yu;
      Pages: 1 - 15
      Abstract: In the last decades, how to detect ships with polarimetric synthetic aperture radar (PolSAR) has become one hot topic. Unfortunately, most of the existing ship detection methods cannot well detect small ships with weak backscattering. To deal with this issue, a ship detection matrix named complete polarimetric covariance matrix [CP] was recently proposed from the perspective of spatial information utilization. Although it is able to improve small ships’ target-to-clutter ratio (TCR) values, its calculation strategy still needs to be rethought due to the possible information loss of some ships. Besides, its mathematical characteristic (i.e., not positive semidefinite) also limits the successful applications of some existing polarimetric theories to it. To overcome these two drawbacks, we here develop an information reconstruction-based polarimetric covariance matrix [IC]. In brief, one new difference calculation strategy is first performed on the Sinclair matrix [ $S$ ], so as to reconstruct its information, by which a feature vector $v$ is subsequently extracted with the Lexicographic matrix basis. Then, via further performing an outer product operation on $v$ , the matrix [IC] is proposed. Meanwhile, to demonstrate the effectiveness of [IC] in ship detection, two different [IC]-based intensity detectors, respectively, named SPANIC and PEDIC, are designed as well. Experiments carried out on three GF-3 PolSAR datasets show that: 1) the proposed matrix [IC] has a better performance than [CP] and the original polarimetric covariance matrix [ $C$ ] in ship detection and 2) compared to the total power detector SPAN and geometrical perturbation-polarimetric notch filter -GP-PNF), both SPANIC and PEDIC can better detect ships, especially the small ships.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • MLE-MPPL: A Maximum Likelihood Estimator for Multipolarimetric Phase
           Linking in MTInSAR

    • Free pre-print version: Loading...

      Authors: Huaping Xu;Guobing Zeng;Wei Liu;Yuan Wang;
      Pages: 1 - 13
      Abstract: Multitemporal synthetic aperture radar interferometry (MTInSAR) is an efficient geodetic tool for Earth surface displacement measurement, and the polarimetric capability of current and upcoming SAR satellites offers a new opportunity to further improve MTInSAR phase series estimation. However, none of the existing estimators for multipolarimetric MTInSAR phase series of distributed scatters (DSs) is derived under the minimum root-mean-square error (RMSE) criterion. In this work, a maximum likelihood estimator for multipolarimetric phase linking (MLE-MPPL) is proposed and the corresponding Cramer–Rao lower bound (CRLB) is also derived by modeling the polarimetric interferometric coherence matrix as the Kronecker product of polarimetric coherence matrix and interferometric coherence matrix. In addition, a new metric called Pol-detR is proposed for the performance evaluation of multipolarimetric MTInSAR phase series estimation in practical scenarios where the RMSE is not feasible any more. The experimental results based on both simulated and real data show that the proposed MLE-MPPL achieves the best estimation performance and is more robust against interchannel interference than existing methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • SAR-TSCC: A Novel Approach for Long Time Series SAR Image Change Detection
           and Pattern Analysis

    • Free pre-print version: Loading...

      Authors: Weisong Li;Peifeng Ma;Haipeng Wang;Chaoyang Fang;
      Pages: 1 - 16
      Abstract: Change detection has played an increasingly important role in multitemporal remote sensing applications recently. Long time series analysis is providing new information of land cover changes and improving the quality and accuracy of the change information being derived from remote sensing. The purpose of this study is to dig for more change temporal information and change pattern information from synthetic aperture radar (SAR) image time series (ITS), which is of great significance for monitoring urban area changes, conducting land use surveys, and renovating illegal constructions. In the study, a novel unified framework for long time series SAR image change detection and change pattern analysis (SAR-TSCC) was proposed for land cover change mapping. To obtain the most notable change time rapidly, a fast SAR ITS change point search method based on pruned exact linear time (SAR-PELT) algorithm was adopted. Meanwhile, the deep time series classification network, named SAR time series transformer (SAR-TST), was implemented to recognize the change patterns, which is based on time series transformer (TST) architecture. Considering the lack of real training data, a novel synthetic data generation method is developed. The combination of the synthetic and real data enhanced the generalization of the classifiers. The proposed framework was used for monitoring a large urbanization area in the northwest of Hong Kong, China. The Cosmo Skymed (CSK) time series data acquired from 2013 to 2020 were exploited for land cover change analysis. Experiment results showed that our approach achieved the state-of-the-art performance, as the time accuracy reached 86% and the classification accuracy on the four main change patterns (impulse, step, cycle, and complex) is over 99%. In particular, the proposed SAR-TST model showed remarkable advantages in the presence of insufficient real data.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Unified Classification Framework for Multipolarization and Dual-Frequency
           SAR

    • Free pre-print version: Loading...

      Authors: Mingliang Liu;Yunkai Deng;Donghong Wang;Xiuqing Liu;Chunle Wang;
      Pages: 1 - 13
      Abstract: For synthetic aperture radar (SAR), multipolarization and multifrequency modes greatly enrich the acquired earth resource information and have been widely applied in remote-sensing fields. In this article, we compare the classification capabilities of multipolarization and dual-frequency SAR. To meet the objective of selecting consistent and complete polarimetric information, a unified classification framework is proposed. In the framework, covariance matrices are used directly as inputs instead of polarimetric indicators. Additionally, the Wishart mixture model (WMM) is utilized to characterize the statistical distribution of polarimetric SAR data. Besides, the data log-likelihood function is utilized to mitigate the influence of the initial values on the expectation-maximization (EM) algorithm. Then, among the combinations of four sample-to-subclass distances and two schemes for obtaining sample-to-class distances, the one with the best classification performance is selected as the default for this framework. In the experiments, the classification capabilities of full polarization (FP), compact polarization (CP), and dual-polarization (DP) modes are first compared through the proposed classification framework. Then, we compare the classification capabilities of dual-frequency SAR in FP, CP, and DP modes. For polarimetric SAR (PolSAR) system design, it is necessary to strike a balance between demand indexes (classification performance, coverage width, and so on) and cost (such as budget, weight, and so on). The comparison results provide a reference for the optimization of polarization modes and frequency bands of the existing multipolarization and dual-frequency SAR payloads and the design of future SAR systems.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • On the Method of Circular Polarimetric SAR Calibration Using Distributed
           Targets

    • Free pre-print version: Loading...

      Authors: Yonghui Han;Pingping Lu;Xiuqing Liu;Wentao Hou;Yao Gao;Weidong Yu;Robert Wang;
      Pages: 1 - 16
      Abstract: The channel imbalance and crosstalks are the two major factors for the polarimetric calibration of the circular quad-polarimetric (CQP) synthetic aperture radar (SAR) systems. In existing methods using distributed targets, the latter is usually ignored, which may lead to unbearable errors in target classification, surface parameter inversion, and so on. To address this issue, this article proposes a modified iterative calibration method using distributed targets that satisfy quasi-azimuth symmetry to calibrate both the channel imbalance and crosstalks of the CQP SAR systems. First, a stable distributed target selection strategy is proposed based on the correlation coefficient of LL polarization and RR polarization, the cross- and co-polarization backscatter ratio, and the equivalent number of looks (ENL). These parameters are insensitive to polarization distortion, and their typical ranges are determined via numerical simulations. Their combination helps select the targets that satisfy the quasi-azimuth symmetry, which is critical for calibrating the crosstalks. Then, the calibration can be conducted using the selected target. Finally, the phase ambiguity of the receive channel imbalance ratio, commonly found in distributed-target-based algorithms, is eliminated using a dipole target. Through the calibration of Gaofen-3 data and the statistical analysis of residual distortion, the effectiveness of the proposed method is verified.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • AMD-HookNet for Glacier Front Segmentation

    • Free pre-print version: Loading...

      Authors: Fei Wu;Nora Gourmelon;Thorsten Seehaus;Jianlin Zhang;Matthias Braun;Andreas Maier;Vincent Christlein;
      Pages: 1 - 12
      Abstract: Knowledge on changes in glacier calving front positions is important for assessing the status of glaciers. Remote sensing imagery provides the ideal database for monitoring calving front positions; however, it is not feasible to perform this task manually for all calving glaciers globally due to time constraints. Deep-learning-based methods have shown great potential for glacier calving front delineation from optical and radar satellite imagery. The calving front is represented as a single thin line between the ocean and the glacier, which makes the task vulnerable to inaccurate predictions. The limited availability of annotated glacier imagery leads to a lack of data diversity (not all possible combinations of different weather conditions, terminus shapes, sensors, etc. are present in the data), which exacerbates the difficulty of accurate segmentation. In this article, we propose attention-multihooking-deep-supervision HookNet (AMD-HookNet), a novel glacier calving front segmentation framework for synthetic aperture radar (SAR) images. The proposed method aims to enhance the feature representation capability through multiple information interactions between low-resolution and high-resolution inputs based on a two-branch U-Net. The attention mechanism, integrated into the two branch U-Net, aims to interact between the corresponding coarse and fine-grained feature maps. This allows the network to automatically adjust feature relationships, resulting in accurate pixel classification predictions. Extensive experiments and comparisons on the challenging glacier segmentation benchmark dataset CaFFe show that our AMD-HookNet achieves a mean distance error (MDE) of 438 m to the ground truth outperforming the current state of the art by 42%, which validates its effectiveness.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Ship Contour Extraction From SAR Images Based on Faster R-CNN and
           Chan–Vese Model

    • Free pre-print version: Loading...

      Authors: Mingda Jiang;Lingjia Gu;Xiaofeng Li;Fang Gao;Tao Jiang;
      Pages: 1 - 14
      Abstract: Compared with most ship detection methods for synthetic aperture radar (SAR) images, ship contour extraction can provide more of the detailed shape and edge information of an observed ship and play a significant role in sea surface monitoring and marine transportation. In this study, a joint ship contour extraction method (faster region convolutional neural network (R-CNN), fast nonlocal mean (FNLM) filter and Chan–Vese model (FFCV) method) was proposed to obtain detailed ship information from SAR images, including ship detection in complex scenes and contour extraction in target slices. First, Faster R-CNN was employed to slice ships from large-scene SAR images. Then, FNLM filtering was applied to denoise and enhance the structural information of the target slices. Finally, an optimized Chan–Vese model was proposed in this article, which can not only accurately extract the contour of the observed ship but also reduce the computation time of the model. The SAR ship detection dataset (SSDD) was selected and finely relabeled to evaluate the contour extraction performance. An evaluation index $R_{N}$ , including quantitative value and offset direction, was developed to evaluate the extraction accuracy of the target contour from the SAR images. Compared with the Mask R-CNN network, the average contour extraction accuracy index $R_{N}$ of the proposed FFCV method reached −0.002 on all the images in the SSDD dataset, and its results were closer to the real ship contours while maintaining the applicability to complex scenes.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Novel Imaging-Based Target Detection and Parameter Estimation Scheme for
           Airborne Multichannel Circular Stripmap SAR

    • Free pre-print version: Loading...

      Authors: Chong Song;Bingnan Wang;Maosheng Xiang;Ruihua Shi;Qinghai Dong;Zhongqi Wang;Yachao Wang;Xiaofan Sun;Lukai Song;
      Pages: 1 - 15
      Abstract: Airborne multichannel circular stripmap synthetic aperture radar (CSSAR)-ground moving target indication (GMTI) has drawn increasing attention in wide-area surveillance, reconnaissance, and traffic monitoring due to its short revisit times. In this article, a novel scheme for CSSAR-GMTI systems is proposed to offer high-resolution focusing of moving targets and enable efficient detection and accurate parameter estimation, which are ensured by integrating moving target imaging with a space–time adaptive processing (STAP) clutter suppression step. The presented scheme develops a target imaging algorithm that forms SAR images focused on a particular 2-D motion parameter to maximize target signal energy and recover its high resolution. Moreover, benefiting from the suitable phase compensation designed in the 2-D frequency domain, moving targets can finally be properly focused in the SAR image without displacement, and their parameters can be uniquely determined. Experiments on an emulated radar dataset are conducted to validate the effectiveness of the proposed scheme.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • An Image-Domain Signal Model for Azimuth Multichannel Reconstruction and
           Its Applications

    • Free pre-print version: Loading...

      Authors: Yuming Jiang;Bing Sun;Jingwen Li;
      Pages: 1 - 13
      Abstract: High-resolution wide-swath (HRWS) synthetic aperture radar (SAR) may benefit from reconstructing signals in the image domain, for taking local characteristics of the scene into account. An image-domain signal model for HRWS SAR to reconstruct nonambiguous images is presented in this article. This model makes the idea of controlling the regional reconstruction performance and taking the advantage of nonuniform scatter distribution possible. Resolution is also preserved perfectly by the proposed model despite whatever tradeoff is made between the azimuth-ambiguity-to-signal ratio (AASR) and the signal-to-noise ratio (SNR). The model is derived on backprojection (BP) images and can deal with different squint angles and beam steering strategies, thanks to the BP algorithm. Although the reconstruction is done pixel by pixel, the computational burden, which depends on beam steering and the specific constraints in reconstruction, may not severely increase. Two methods for Spotlight SAR and two methods for Stripmap SAR are also presented based on the proposed model to verify the model and demonstrate its potential. Simulations on both the point target and extended target with different squint angles are carried out to verify the proposed methods, which also verify the proposed model indirectly.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Adaptive Self-Supervised SAR Image Registration With Modifications of
           Alignment Transformation

    • Free pre-print version: Loading...

      Authors: Shasha Mao;Jinyuan Yang;Shuiping Gou;Kai Lu;Licheng Jiao;Tao Xiong;Lin Xiong;
      Pages: 1 - 15
      Abstract: Considering that deep learning achieves the prominent performance, it has been applied to synthetic aperture radar (SAR) image registration to improve the registration accuracy. In most methods, a deep registration model is constructed to classify matched points and unmatched points, in which SAR image registration is regarded as a supervised two-classification problem. However, it is difficult to annotate massive matched points manually in practice, which limits the performance of deep networks. Besides, inevitable differences among SAR images easily cause that some training and testing samples are inconsistent, which probably brings negative effects for training a robust registration model. To address these problems, we propose an adaptive self-supervised SAR image registration method, where SAR image registration is regarded as a self-supervised task rather than the supervised two-classification task. Inspired by self-supervised learning, we consider each point on SAR images as a category-independent instance, which mitigates the requirement of manual annotations. Based on key points from images, a self-supervised model is constructed to explore the latent feature of each key point, and then, pairs of match points are sought via evaluating similarities among key points and used to calculate the alignment transformation matrix. Meanwhile, to enhance the consistency of samples, we design a new strategy that constructs multiscale samples by transforming key points from one image into another, which avoids inevitable diversities between two images effectively. In particular, the constructed samples feeding to the self-supervised model are adaptively updated with the modification of the transformation matrix in iterations. Moreover, the similarity of maximal public areas (MPAS) indicator is proposed to assist in estimating the transformation. Finally, experimental results illustrate that the proposed method achieves more accurate registrations than other compared met-ods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Classification Matters More: Global Instance Contrast for Fine-Grained SAR
           Aircraft Detection

    • Free pre-print version: Loading...

      Authors: Danpei Zhao;Ziqiang Chen;Yue Gao;Zhenwei Shi;
      Pages: 1 - 15
      Abstract: Since significant intraclass differences and inconspicuous interclass variations, fine-grained aircraft detection in synthetic aperture radar (SAR) images is challenging. Also, the inherent lack of detailed features and severe noise interference in SAR images make it difficult to learn class-specific feature representations. Current detection approaches focus more on localization accuracy and ignore classification performance, which is more critical in fine-grained detection. To address the above challenges, we present GICNet: global instance contrast (GIC) for fine-grained SAR aircraft detection a global instance-level contrast module is proposed to improve interclass divergences and intraclass compactness. With a specially constructed global instance set, GICNet can contrast a large number of different aircraft targets while keeping a small batch size. Furthermore, we design a novel quality-aware focal loss (QAFL) to facilitate the accurate classification of well-localized aircraft targets. Meanwhile, to maintain localization performance, we develop a new edge-aware bounding-box refinement (EABR) module to refine predicted coarse bounding boxes. Experimental results show that our GICNet outperforms current advanced detectors and achieves a new state-of-the-art performance on the GaoFen-3 SAR aircraft detection dataset. In particular, GICNet also has advantages in reducing misclassification and recognizing well-located targets.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • HRLE-SARDet: A Lightweight SAR Target Detection Algorithm Based on Hybrid
           Representation Learning Enhancement

    • Free pre-print version: Loading...

      Authors: Zheng Zhou;Jie Chen;Zhixiang Huang;Jianming Lv;Jiaxing Song;Honglin Luo;Bocai Wu;Yingsong Li;Paulo S. R. Diniz;
      Pages: 1 - 22
      Abstract: In recent years, deep learning has been widely used in remote sensing, especially in the field of synthetic aperture radar (SAR) image target detection. However, all of these deep learning models continue increasing the network’s depth and width without maintaining a good balance between accuracy and speed. Therefore, in this article, we propose a hybrid representation learning-enhanced SAR target detection algorithm based on the unique features of SAR images from a lightweight perspective called HRLE-SARDet. First, we design a lightweight and scattering feature extraction backbone that is more suitable for SAR image data. Second, for the multiscale feature discrepancy, we design a new multiscale feature fusion neck. Next, to better extract the scattering information from small targets of SAR images and improve the detection accuracy, we design a lightweight hybrid representation learning enhancement module. Finally, to better fit target detection for SAR image datasets, we redesign a more flexible loss function, which allows for an easy adjustment of the importance of polynomial bases according to the target task and dataset. Extensive experimental results on three SAR image ship target datasets (SSDD, AIR-SARShip-2.0, and HRSID) and a newly released large multiclass target SAR dataset (MSAR-1.0) show that our HRLE-SARDet achieves 98.4%, 79.2%, 92.5%, and 88.4% mean average precision (mAP) with only 1.09 M parameters and 2.5 G floating-point operations (FLOPs) on the SSDD, AIR-SARShip-2.0, HRSID, and MSAR-1.0 datasets, respectively, which is an excellent performance.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Integrated Detection and Imaging Algorithm for Radar Sparse Targets via
           CFAR-ADMM

    • Free pre-print version: Loading...

      Authors: Pucheng Li;Zegang Ding;Tianyi Zhang;Yangkai Wei;Yongpeng Gao;
      Pages: 1 - 15
      Abstract: Most research on sparsity-driven synthetic aperture radar (SAR) imaging has been carried out in $ell _{1}$ -norm regularization and considers that the SAR image contains only targets and noise, which ignores the clutter and seriously degrades classical algorithms. To address this problem, we propose an integrated detection and imaging algorithm for radar sparse targets with constant false alarm rate (CFAR) regularization by alternating direction method of multipliers (ADMM), called CFAR-ADMM, and we further introduce total variation (TV) regularization and propose the more robust CFAR-TV-ADMM. First, a more complete echo signal model, which considers targets, the clutter, and the noise simultaneously, is established. Then, inspired by the CFAR detection, a novel regularization with sparse target awareness is proposed. The proposed regularization can obtain the statistical characteristics of clutter and noise region by region, and distinguish whether the current cell contains the target effectively and accurately. Benefiting from this novel regularization, CFAR-ADMM and TV-CFAR-ADMM can not only realize the sparse imaging but also detect sparse targets simultaneously, which can reduce the propagation error caused by cascading processing and improve the solution accuracy. Finally, the proposed algorithm is verified by simulation data results, phase transition analysis, and real data experiments.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Few-Shot Class-Incremental SAR Target Recognition Based on Hierarchical
           Embedding and Incremental Evolutionary Network

    • Free pre-print version: Loading...

      Authors: Li Wang;Xinyao Yang;Haoyue Tan;Xueru Bai;Feng Zhou;
      Pages: 1 - 11
      Abstract: It is difficult to realize effective synthetic aperture radar (SAR) automatic target recognition (ATR) in open scenarios because the ATR model cannot continuously learn from new classes with limited training samples. When adding new classes to the previously trained model, the capability of recognizing old classes may lose due to severe overfitting. To tackle this problem, a few-shot class-incremental SAR ATR method, namely, hierarchical embedding and incremental evolutionary network (HEIEN), is proposed in this article. First, a hierarchical embedding network and a hybrid distance-based classifier are constructed for basic feature extraction and classification. Then, in order to obtain more accurate decision boundaries, an adaptive class-incremental learning (ACIL) module is designed to adjust the weights of classifiers in all tasks by collecting context information from the past to the present. Finally, a pseudo-incremental training strategy is designed to enable effective model training with only a few samples. Experimental results on the moving and stationary target acquisition and recognition (MSTAR) benchmark data set have illustrated that HEIEN performs well with remarkable advantages in few-shot class-incremental SAR ATR tasks.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Modeling the Effects of Oscillator Phase Noise and Synchronization on
           Multistatic SAR Tomography

    • Free pre-print version: Loading...

      Authors: Eric Loria;Samuel Prager;Ilgin Seker;Razi Ahmed;Brian Hawkins;Marco Lavalle;
      Pages: 1 - 12
      Abstract: Recent results have highlighted the potential ability of bistatic and multistatic synthetic aperture radar (SAR) tomographers to measure vegetation structure and surface topography. However, the quality of SAR tomographic measurements with multiple platforms is impacted by the phase instability in each platform’s oscillator. The phase noise, if uncompensated, may lead to degradation in the SAR data products such as increased sidelobe levels, reduced peak amplitude of the impulse response, and low-frequency phase modulation, among others. In this work, we model and examine the effects of oscillator phase noise on tomographic SAR signals for spaceborne missions flying in formation. A synchronization process is also adopted to help mitigate oscillator phase errors by measuring and predicting relative phase offsets at prescribed temporal intervals. A simulation tool was developed to examine the point target response (PTR) as seen by realistic satellite constellations in low Earth orbit using different quality oscillators, radar configurations, and synchronization configurations. A first analysis of a multiplatform tomographic SAR mission suggests that a system without a dedicated physical link with minimal effects on the PTR may be achievable using current oscillators. Our analysis also shows that phase noise has differing effects on multistatic radar modes. Tomograms formed with a system operating in single-input–multiple-output (SIMO) mode are the most affected by an oscillator phase noise error, followed by multiple-input–multiple-output (MIMO), with negligible effects on the single-input single-output (SAR-SISO) mode. These trade studies and the simulation tool can be used to help inform the design of future multistatic radar missions.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Practical Issue Analyses and Imaging Approach for Hypersonic Vehicle-Borne
           SAR With Near-Vertical Diving Trajectory

    • Free pre-print version: Loading...

      Authors: Shiyang Tang;Xintian Zhang;Zixuan He;Zhanye Chen;Wangwang Du;Yinan Li;Juan Zhang;Ping Guo;Linrang Zhang;Hing Cheung So;
      Pages: 1 - 16
      Abstract: As a frontier technology in radar imaging, hypersonic vehicle (HSV)-borne synthetic aperture radar (SAR) has several practical issues to be dealt with, namely, ground resolution capability, pulse repetition frequency (PRF) selection, and beam pointing description, especially for the near-vertical diving trajectory because of the extremely small angle between the velocity and slant range vectors. Moreover, its focusing approach design is greatly challenged by very large cross-couplings and spatial variations. Considering these practical problems, the constraints between system performance and parameter selection are analyzed first to obtain the parameter optimization procedure and avoid system design deviation. Then, a frequency radius/angle algorithm (FRAA) is devised, which is an extension of the radius/angle algorithm (RAA) performed in the 2-D frequency domain. In the FRAA, the new range equation and 2-D frequency interpolation function are reconstructed with high accuracy by quadratic fitting and 3-D expansion. Compared with RAA, FRAA is more suitable for the HSV SAR with near-vertical diving trajectory. Simulation results verify the effectiveness of the proposed approach.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • MT-InSAR Unveils Dynamic Permafrost Disturbances in Hoh Xil (Kekexili) on
           the Tibetan Plateau Hinterland

    • Free pre-print version: Loading...

      Authors: Ping Lu;Jiangping Han;Yonghong Yi;Tong Hao;Fujun Zhou;Xianglian Meng;Yinsheng Zhang;Rongxing Li;
      Pages: 1 - 16
      Abstract: Hoh Xil is an uninhabited extremity secluded on the Tibetan Plateau hinterland. A complete mapping of ground motion variation in Hoh Xil is essential for an in-depth understanding of the terrain’s responses to climate change on the Tibetan Plateau. However, the inaccessibility and extremely harsh environment impeded extensive field investigations on landform alteration and its formative process. Such difficulty can be resolved by interferometric synthetic aperture radar (InSAR), which enables a broad detection of subtle permafrost motions at millimeter precision. This study, for the first time, accomplished a multitemporal InSAR (MT-InSAR) deformation mapping from 2015 to 2020 in Hoh Xil, with a wide coverage of about 200 000 km2. 1592 Sentinel-1 images were processed based on the small baseline subset (SBAS) technique. The results show that Hoh Xil was experiencing dynamic permafrost disturbances. Thawing permafrost with both a linear subsidence rate higher than 2 mm/year and a periodic amplitude over 2 mm was primarily detected in areas of flat or gentle slopes. The InSAR cumulative deformation is highly correlated with permafrost thawing depth. Significant lag times were identified between seasonal oscillation of InSAR deformation and land surface temperature (LST). Thermokarst landforms of retrogressive thaw slumps and thermokarst lakes broadly formed and dynamically evolved as a consequence of permafrost degradation. Particularly, widespread thawing permafrost characterized by the spatial clustering of thermokarst lakes appeared to occur in areas adjacent to large lakes. The discovered dynamic permafrost disturbances in Hoh Xil manifested even the secluded Tibetan Plateau hinterland was facing the threat of climate change.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Modified Range Model and Extended Omega-K Algorithm for
           High-Speed-High-Squint SAR With Curved Trajectory

    • Free pre-print version: Loading...

      Authors: Tinghao Zhang;Yachao Li;Jun Wang;Mengdao Xing;Liang Guo;Peng Zhang;
      Pages: 1 - 15
      Abstract: An accurate range model with acceleration, the coupling phase terms, and spatial-variant (SV) Doppler parameters are the main issues to be solved in high-speed-high-squint synthetic aperture radar (HSHS-SAR) with a curved trajectory. For these issues, an extended Omega-K (EOK) algorithm is developed in this article. The proposed EOK algorithm mainly includes the following four aspects. First, a modified range model (AMRM) considering 3-D acceleration for a curved trajectory is established. Then, the coupling between the range and azimuth direction is removed by the modified Stolt mapping (MSM). Subsequently, an improved high-order SV phase correction approach is derived to eliminate the azimuth dependence of Doppler parameters. Finally, in order to avoid zeros-padding operation, the proposed method focuses on the sub-aperture data in the range time and azimuth frequency domain through data aligning processing. The experimental results of both simulation and real data verify the effectiveness of the proposed method.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • PAN: Part Attention Network Integrating Electromagnetic Characteristics
           for Interpretable SAR Vehicle Target Recognition

    • Free pre-print version: Loading...

      Authors: Sijia Feng;Kefeng Ji;Fulai Wang;Linbin Zhang;Xiaojie Ma;Gangyao Kuang;
      Pages: 1 - 17
      Abstract: Machine learning methods for synthetic aperture radar (SAR) image automatic target recognition (ATR) can be divided into two main types: traditional methods and deep learning methods. The deep learning methods can learn the high-dimensional features of the target directly and usually obtain high target recognition accuracy. However, they lack full consideration of SAR targets’ inherent characteristics, resulting in poor generalization and interpretation ability. Compared with deep learning methods, traditional methods can get more interpretable and stable results with model-based features. In order to take full advantage of these two kinds of methods, we propose the target part attention network (PAN) based on the attributed scattering center (ASC) model to integrate the electromagnetic characteristics with the deep learning framework. First, considering the importance of scattering structure for SAR ATR, we design a target part model based on the ASC model. Then, a novel part attention module based on the scaled dot-product attention mechanism is proposed, which directly associates the features of target parts with the classification results. Finally, we give the derivation method of the importance of each part, which is of great significance for practical application and the interpretation of SAR ATR. Experiments on the MSTAR dataset demonstrate the effectiveness of the proposed PAN. Compared with existing studies, it can achieve higher and more robust classification accuracy under different complex conditions. Furthermore, combined with the importance of parts, we constructed two effective interpretable analysis methods for deep learning network classification results.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Permittivity Extraction of Soil Samples Using Coaxial-Line Measurements by
           a Simple Calibration

    • Free pre-print version: Loading...

      Authors: Hafize Hasar;Ugur C. Hasar;Hamdullah Ozturk;Mucahit Izginli;Nevzat Aslan;Taskin Oztas;Mehmet Ertugrul;Mehmet Karayilan;Omar M. Ramahi;
      Pages: 1 - 8
      Abstract: A new microwave method is proposed for accurate determination of complex permittivity $varepsilon _{textrm {rs}} = varepsilon _{textrm {rs}}^{prime }--j varepsilon _{textrm {rs}}^{prime prime }$ of soil samples inserted over a holder within the Electronic Industries Association (EIA) 1-5/8” coaxial measurement cell. Such a determination could be indirectly correlated with the volumetric moisture content of soil samples by microwave measurements. The method has three main advantages. First, it utilizes a simple calibration procedure involving uncalibrated measurements of an empty cell, the same cell loaded with a soil holder (a dielectric sample), and the same cell with a soil sample over this holder, thus eliminating the need for any formal calibration procedure. Second, it uses one measurement cell for extracting $varepsilon _{textrm {rs}}$ . Third, it does not require any numerical toolbox for determining $varepsilon _{textrm {rs}}$ . The method is validated by simulations of a synthesized soil sample and by experiments with a low-loss polyethylene sample. Its accuracy is examined in reference to: 1) two measurement cells with different lengths (length independence); 2) the position of the holder in the cell; and 3) an offset in sample length. Calibration curves (the volumetric moisture content $theta _{V}$ versus $varepsilon _{textrm {rs}}^{prime }$ ) obtained from fitting measured $varepsilon _{textrm {rs}}^{prime }$ by our method to $theta _{V}$ at 2 and 3 GHz are compared with other calibration curves in the literature for the analysis of the performance of the proposed method (PM). It is shown that calibration curves obtained from our method are similar to those obtained from other methods requiring complex calibration procedures.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • An Antenna Pattern Correction Algorithm for Conical Scanning Spaceborne
           Radiometers: The CIMR Case

    • Free pre-print version: Loading...

      Authors: Alessandro Lapini;Ada Vittoria Bosisio;Giovanni Macelloni;Marco Brogioni;
      Pages: 1 - 15
      Abstract: The rapid evolution of the effects observed in various areas of our planet related to climate change poses urgent questions about the knowledge of the state of the polar area and requires satellite acquisitions with fine spatial resolution and high accuracy to develop advanced products. The Copernicus Imaging Microwave Radiometer (CIMR) mission, based on a multifrequency microwave radiometer and designed to observe the ocean, sea ice, and Arctic environment, requires brightness temperature measurements with a total absolute uncertainty of 0.5 K and a spatial resolution of 5 km. This constraint demands very large reflectors with a gain value of tens of decibels. Mechanical constraints will be attained by using a mesh reflector, which guarantees the required resolution but with the drawback of a radiation pattern characterized by many grating lobes that contaminate the value of the brightness temperature associated with the boresight position. In this article, an antenna pattern correction (APC) is proposed to correct these effects. The algorithm takes advantage of an iterative formulation based on the Jacobi Method, providing a suitable correction that depends on the chosen spatial resolution. The APC algorithm was tested at both K- and Ka-bands with similar performance. Here, only the results from the latter are shown, as its antenna pattern is the most challenging among CIMR.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Toward an Improved Surface Roughness Parameterization Model for Soil
           Moisture Retrieval in Road Construction

    • Free pre-print version: Loading...

      Authors: Thi Mai Nguyen;Jeffrey P. Walker;Nan Ye;Jayantha Kodikara;
      Pages: 1 - 13
      Abstract: In passive microwave remote sensing, the estimation of the surface roughness parameter is a significant obstacle for soil moisture (SM) retrieval. For a given SM content, the geometric soil surface roughness has been shown to have a large impact on the surface emission at L-band frequency, which affects the SM retrieval success when using the information observed from the radiometer and is represented through the so-called the surface roughness parameter ( $H_{R}$ ). Moreover, no previous study has examined the effect of this factor in the context of road construction, where the geometric soil surface roughness is affected by the compaction process, resulting in a substantial change in roughness before and after compaction. Accordingly, a series of experiments at various compaction levels and SM contents was performed for a sand subgrade material in order to identify their effects on $H_{R}$ . The soil brightness temperature (TB) was measured using an L-band radiometer at different incidence angles and a laser profiler was used to measure the surface roughness standard deviation ( $sigma$ ) before and after compaction. The results of this article have demonstrated that the incidence angle ( $theta$ ) and SM both affect $H_{R}$ and its relation to the geometric soil surface roughness. Importantly, these factors are not accounted for by existing models. Consequently, a modified surface roughness parameter ( $H_{R}$ ) model, based on the traditional Choudhury model, was developed to include the contribution of these two factors, and its impact o- the accuracy of SM retrieval results tested. Specifically, it was shown that it is possible to obtain SM retrieval results with an accuracy of 0.04 cm3/cm3 at almost all incidence angles using either dual-polarization [both horizontal (H) and vertical polarization (V)] or only vertical polarization observations. The modified surface roughness parameter ( $H_{R}$ ) model has improved the performance of the SM retrieval model to achieve an accuracy of 0.04 cm3/cm3, whereas the traditional Choudhury model achieved an accuracy of only 0.05 cm3/cm3.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Registration and Fusion Using Reflection in Passive Millimeter-Wave Images

    • Free pre-print version: Loading...

      Authors: Wenning Xu;Liang Lang;Liangqi Gui;Kaixin Deng;Fangzhou Tang;
      Pages: 1 - 9
      Abstract: Passive millimeter-wave (PMMW) imaging technology is widely used in civilian and military applications. However, there are reflections similar to the optical band in PMMW images, which have a negative influence on target detection and recognition. In this article, we present a reflection-based method to enhance the target features in PMMW images. The dividing line between the target and reflection is obtained by the similarity of brightness temperature (TB). By combining the similarity and reflection principle, we propose a new method to obtain the feature points of the target and reflection for registration. Then, the weighted method based on region TB is used to fusion target and reflection. Finally, to avoid interference with target detection and recognition, the reflection is removed. The experimental results show that the method can obtain higher contrast and more accurate target information.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • RFI Localization Using Jointly Non-Convex Low-Rank Approximation and
           Expanded Virtual Array in Microwave Interferometric Radiometry

    • Free pre-print version: Loading...

      Authors: Yanyu Xu;Dong Zhu;Fei Hu;Peng Fu;
      Pages: 1 - 15
      Abstract: The scientific goal of the Soil Moisture and Ocean Salinity (SMOS) mission is to retrieve the geophysical parameter from brightness temperature (TB) maps. However, radio frequency interference (RFI) significantly influences the interpretation of TB maps, leading to a deteriorated retrieval performance. RFI localization is essential for switching off these illegal emitters and mitigating their impacts on TB maps. This article proposes a novel high-resolution RFI localization method via jointly non-convex low-rank approximation and expanded virtual array (EVA). Concretely, the RFI localization problem is first formulated from the perspective of non-convex low-rank recovery, which better approximates the rank of the covariance matrix collecting visibility samples. Then, we propose the EVA concept by relaxing the size constraint on the physical antenna array. Moreover, we use a new algorithm based on the joint Schatten- $p$ and $Lp$ (JSL) norms to solve the above non-convex low-rank recovery problem. This JSL algorithm can improve the spatial resolution for RFI localization. Combining the JSL algorithm and the EVA can further improve the detection performance and enhance the spatial resolution for RFI localization. The experimental results using synthetic data and real SMOS data prove that the proposed method shows enhanced spatial resolution, better detection performance, and competitive or better localization accuracy compared with the currently existing methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Generation of Spatial-Seamless AMSR2 Land Surface Temperature in China
           During 2012–2020 Using a Deep Neural Network

    • Free pre-print version: Loading...

      Authors: Yihua Lian;Si-Bo Duan;Cheng Huang;Wenjing Han;Meng Liu;
      Pages: 1 - 18
      Abstract: Land surface temperature (LST) reflects the cold and hot conditions of the land surface and is one of the most important geophysical parameters in the study and research of the land–atmosphere system. Passive microwave (PMW) is one of the primary techniques for obtaining spatially continuous LST at regional, continental, and global scales. However, there is an orbital gap in the LST retrieved from PMW (PMW LST) due to the scanning scheme of the PMW sensor, which limits the application of PMW LST, so it is necessary for the proposed some methods to fill the orbital gap of PMW LST. In this study, a new orbital gap-filling method based on a deep neural network (DNN) was developed to address the issue of PMW LST orbital gaps. This method first established the DNN model based on the nonlinear relationship between AMSR2 LST and 11 environmental variables and then used the DNN model to generate a new spatially continuous LST product, namely, DNN-LST, and, finally, used DNN-LST to fill the orbital gaps of AMSR2 LST to generate the daytime/nighttime spatially seamless gap-filled LST (GF-LST) product for China from 2012 to 2020. GF-LST can more correctly represent the spatiotemporal variation of surface temperature in China than AMSR2 LST because it has continuous spatial texture information and no obvious boundary reconstruction effect. After verifying the accuracy of GF-LST products through simulated gap region validation and in situ validation, it can be found that: 1) DNN-LST in simulated gap regions showed high accuracy during the daytime and nighttime on July 15, 2012–2020, and the mean values of bias and root mean square error (RMSE) compared with AMSR2 LST at day (night) were, respectively, −0.08 K (−0.22 K) and 1.89 K (2.23 K); 2) the accuracy of DNN-LST was the best in autumn (mean RMSE values of 1.43 K at day and 1.89 K at night) and the worst in winter (mean RMSE values of 2.3- K at day and 2.36 K at night), no matter during daytime or nighttime, in different seasons in 2015–2017; 3) the RMSE value of DNN-LST during nighttime was slightly higher than the RMSE value of DNN-LST during daytime; and 4) the accuracy of DNN-LST was equivalent to AMSR2 LST, that is, the unbiased RMSE (ubRMSE) of DNN-LST and AMSR2 LST was all about 4 K compared with in situ LST, but the ubRMSE of DNN-LST was slightly lower than AMSR2 LST. The above accuracy validation analysis shows that DNN-LST has good robustness and good spatial consistency with AMSR2 LST and can be well used to fill the orbital gap of AMSR2 LST to generate spatial seamless GF-LST product.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Deep Learning Imaging for 1-D Aperture Synthesis Radiometers

    • Free pre-print version: Loading...

      Authors: Haofeng Dou;Chengwang Xiao;Hao Li;Yinan Li;Pengju Dang;Rongchuan Lv;Pengfei Li;Guangnan Song;Yuanchao Wu;Xiaojiao Yang;Shubo Liu;Renzhi Jiang;Wenjing Wang;
      Pages: 1 - 16
      Abstract: For 1-D aperture synthesis (1-D AS) radiometers, truncated sampling occurs in the frequency domain due to the system baseline limitation. Therefore, there is an obvious Gibbs oscillation in the reconstructed image. To solve this problem, an imaging method based on a 1-D convolutional neural network (1-D CNN) is proposed in this article. Compared with deep learning methods based on 2-D convolutions, the 1-D convolution not only reduces the amount of computation but also produces further performance improvements. The input data of the network are the 1-D visibility function samples, and the output data are the 1-D brightness temperature (BT) samples. The network learns the mapping relationship from the training of the 1-D visibility function samples and 1-D BT samples to complete 1-D AS imaging without any prior knowledge. To verify the performance of this imaging method, simulations and experiments based on the airborne C-band 1-D microwave interferometric radiometer (ACMIR) system are implemented. The simulation and experimental results demonstrate that the proposed AS-CNN method achieves higher performance than the inverse fast Fourier transform (IFFT) method in terms of image quality and Gibbs phenomenon suppression. In the case of an antenna failure and missing baseline, the AS-CNN method proposed in this article can still obtain a BT image with high imaging quality, which shows that the robustness of the network is better than that of the IFFT method.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Image Reconstruction of Synthetic Aperture Radiometer by Transformer

    • Free pre-print version: Loading...

      Authors: Chengwang Xiao;Haofeng Dou;Hao Li;Rong Jin;Ren Zhai;Wenjing Wang;Rongchuan Lv;Yinan Li;
      Pages: 1 - 15
      Abstract: In passive microwave remote sensing of the Earth, compared with real aperture radiometer, synthetic aperture radiometer (ASR) is a very powerful instrument with many advantages. However, the system design is more complex than the real aperture radiometer, and the hardware ideality is often not guaranteed. The nonideal characteristics of the system hardware will bring a variety of errors to the system, which will cause the Fourier transform relationship between the visibility function and the brightness temperature image to no longer be established, thus reducing the quality of microwave brightness temperature image reconstruction by traditional methods. In this article, a new microwave brightness temperature image reconstruction method for ASR by transformer is proposed. This method uses a specially designed transformer structure to extract the spectrum features in the visibility function. This method learns the mapping relationship between the visibility function and the original scene brightness temperature image through the supervised learning method, and learns as much as possible the spectrum information contained in the original scene brightness temperature image. Moreover, when there are missing baselines, this method will supplement the missing observation frequency information, so as to obtain better reconstructed image quality. With the above-mentioned advantages, this method can suppress the Gibbs oscillation, and greatly reduce the sidelobe. Compared with the existing reconstruction methods, whether missing baselines or not, the proposed image reconstruction method by transformer has advantages in image quality. We verify the performance of this brightness temperature image reconstruction method through simulation and experiment.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Blind Superresolution of Satellite Videos by Ghost Module-Based
           Convolutional Networks

    • Free pre-print version: Loading...

      Authors: Zhi He;Dan He;Xiaofang Li;Rongning Qu;
      Pages: 1 - 19
      Abstract: Deep learning (DL)-based video satellite superresolution (SR) methods have recently yielded superior performance over traditional model-based methods by using an end-to-end manner. Existing DL-based methods usually assume that the blur kernels are known and, thus, do not model the blur kernels during restoration. However, this assumption is rarely held for real satellite videos and leads to oversmoothed results. In this article, we propose a Ghost module-based convolution network model for blind SR of satellite videos. The proposed Ghost module-based video SR (GVSR) method, which assumes that the blur kernel is unknown, consists of two main modules, i.e., the preliminary image generation module and the SR results’ reconstruction module. First, the motion information from adjacent video frames and the wrapped images are explored by an optical flow estimation network, the blur kernel is flexibly obtained by a blur kernel estimation network, and the preliminary high-resolution image is generated by feeding both blur kernel and wrapped images. Second, a reconstruction network consisting of three paths with attention-based Ghost (AG) bottlenecks is designed to remove artifacts in the preliminary image and obtain the final high-quality SR results. Experiments conducted on Jilin-1 and OVS-1 satellite videos demonstrate that the qualitative and quantitative performance of our proposed method is superior to current state-of-the-art methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Spectral Super-Resolution Based on Dictionary Optimization Learning via
           Spectral Library

    • Free pre-print version: Loading...

      Authors: Hao-Fang Yan;Yong-Qiang Zhao;Jonathan Cheung-Wai Chan;Seong G Kong;
      Pages: 1 - 16
      Abstract: Extensive works have been reported in hyperspectral images (HSIs) and multispectral images (MSIs) fusion to raise the spatial resolution of HSIs. However, the limited acquisition of HSIs has been an obstacle to such approaches. Spectral super-resolution (SSR) of MSI is a challenging and less investigated topic, which can also provide high-resolution synthetic HSIs. To deal with this high ill-posedness problem, we perform super-resolution enhancement of MSIs in the spectral domain by incorporating a spectral library as a priori. First, an aligned spectral library, which maps the open-source spectral library to a specific spectral library created for the reconstructed HR HSI, is represented. An intermediate latent HSI is obtained by fusing the spatial information from MSI and the hyperspectral information from a specific spectral library. Then, we use low-rank attribute embedding to transfer latent HSI into a robust subspace. Finally, a low-rank HSI dictionary representing the hyperspectral information is learned from the latent HSI. The adaptive sparse coefficient of MSI is obtained with a nonnegative constraint. By fusing these two terms, we get the final HR HSI. The proposed SSR model does not require any pretraining stages. We confirm the validity and superiority of our proposed SSR algorithm by comparing it with several benchmark state-of-the-art approaches on different datasets.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Modality-Aware Feature Integration for Pan-Sharpening

    • Free pre-print version: Loading...

      Authors: Man Zhou;Jie Huang;Feng Zhao;Danfeng Hong;
      Pages: 1 - 12
      Abstract: Pan-sharpening aims to super-solve low-spatial resolution multiple spectral (MS) images with the guidance of high-resolution (HR) texture-rich panchromatic (PAN) images. Recently, deep-learning-based pan-sharpening approaches have dominated this field and achieved remarkable advancement. However, most promising algorithms are devised in one-way mapping and have not fully explored the mutual dependencies between PAN and MS modalities, thus impacting the model performance. To address this issue, we propose a novel information compensation and integration network for pan-sharpening by effective cross-modality joint learning in this work. First, the cross-central difference convolution is employed to explicitly extract the texture details of the PAN images. Second, we implement the compensation process by imitating the classical back-projection (BP) technique where the extracted PAN textures are employed to guide the intrinsic information learning of MS images iteratively. Subsequently, we devise the hierarchical transformer to integrate the comprehensive relations of stage-iteration information from spatial and temporal contexts. Extensive experiments over multiple satellite datasets demonstrate the superiority of our method to the existing state-of-the-art methods. The source code is available at https://github.com/manman1995/pansharpening.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Mutiscale Hybrid Attention Transformer for Remote Sensing Image
           Pansharpening

    • Free pre-print version: Loading...

      Authors: Wengang Zhu;Jinjiang Li;Zhiyong An;Zhen Hua;
      Pages: 1 - 16
      Abstract: Pansharpening methods play a crucial role for remote sensing image processing. The existing pansharpening methods, in general, have the problems of spectral distortion and lack of spatial detail information. To mitigate these problems, we propose a multiscale hybrid attention Transformer pansharpening network (MHATP-Net). In the proposed network, the shallow feature (SF) is first acquired through an SF extraction module (SFEM), which contains the convolutional block attention module (CBAM) and dynamic convolution blocks. The CBAM in this module can filter initial information roughly, and the dynamic convolution blocks can enrich the SF information. Then, the multiscale Transformer module is used to obtain multiencoding feature images. We propose a hybrid attention module (HAM) in the multiscale feature recovery module to effectively address the balance between the spectral feature retention and the spatial feature recovery. In the training process, we use deep semantic statistics matching (D2SM) loss to optimize the output model. We have conducted extensive experiments on several known datasets, and the results show that this article has good performance compared with other state of the art (SOTA) methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Unsupervised Domain Adaptation Augmented by Mutually Boosted Attention for
           Semantic Segmentation of VHR Remote Sensing Images

    • Free pre-print version: Loading...

      Authors: Xianping Ma;Xiaokang Zhang;Zhiguo Wang;Man-On Pun;
      Pages: 1 - 15
      Abstract: This work investigates unsupervised domain adaptation (UDA)-based semantic segmentation of very high-resolution (VHR) remote sensing (RS) images from different domains. Most existing UDA methods resort to generative adversarial networks (GANs) to cope with the domain shift problem caused by the discrepancies across different domains. However, these GAN-based UDA methods directly align two domains in the appearance, latent, or output space based on convolutional neural networks (CNNs), making them ineffective in exploiting long-range dependencies across the high-level feature maps derived from different domains. Unfortunately, such high-level features play an essential role in characterizing RS images with complex content. To circumvent this obstacle, a mutually boosted attention transformer (MBATrans) is proposed to capture cross-domain dependencies of semantic feature representations in this work. Compared with conventional UDA methods, MBATrans can significantly reduce domain discrepancies by capturing transferable features using global attention. More specifically, MBATrans utilizes a novel mutually boosted attention (MBA) module to align cross-domain feature maps while enhancing domain-general features. Furthermore, a novel GAN-based network with improved discriminative capability is devised by integrating an additional discriminator to learn domain-specific features. Extensive experiments on two large-scale VHR RS datasets, namely, International Society for Photogrammetry and Remote Sensing (ISPRS) Potsdam and Vaihingen, confirm the superior performance of the proposed MBATrans-augmented GAN (MBATA-GAN) architecture. The source code in this work is available at https://github.com/sstary/SSRS.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Pansharpening Method Based on Hybrid-Scale Estimation of Injection Gains

    • Free pre-print version: Loading...

      Authors: Yan Shi;Aiyong Tan;Na Liu;Wei Li;Ran Tao;Jocelyn Chanussot;
      Pages: 1 - 15
      Abstract: The injection scheme provides an efficient way for CS- and MRA-based pansharpening approaches. Within this paradigm, the estimation of injection gains is one of the keys to pansharpening outcomes, which has attracted much attention in the community. Most of the existing models are derived from the regression methodology. Hence, the reference is indispensable for the estimation. However, the reference is unavailable in practice, and therefore, the estimation is usually performed at a degraded scale. This article is devoted to the estimation of injection gains without reference. A hybrid-scale (HS) estimation, which involves both the high-resolution and low-resolution data, is proposed, along with three HS models. The proposed method features a context-based and fast implementation with fewer tunable parameters. Experimental results show that the HS models yield more accurate and robust results compared with the typical regression-based models, and they are also competitive with the state-of-the-art approaches.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • DWin-HRFormer: A High-Resolution Transformer Model With Directional
           Windows for Semantic Segmentation of Urban Construction Land

    • Free pre-print version: Loading...

      Authors: Zhen Zhang;Xin Huang;Jiayi Li;
      Pages: 1 - 14
      Abstract: In this article, a deep neural network for semantic segmentation of high-resolution remote sensing images is proposed for urban construction land classification. The network follows a high-resolution network (HRNet) architecture. Specifically, a directional self-attention on the paths of different resolutions is proposed, aiming to correct the directional bias caused by the attention of strip windows during the model learning, while also reducing the computational complexity, and allowing the model to improve both the accuracy and the speed. At the end of the network, a distributed alignment module with spatial information is constructed to train additional learnable parameters, to adjust the biased decision boundaries through a two-stage learning strategy, and alleviate the problem of accuracy degradation due to the unbalanced training data. We tested the proposed method and compared it with the current state-of-the-art (SOTA) semantic segmentation methods on the Luojia-fine-grained land cover (FGLC) dataset and the Wuhan Dense Labeling Dataset (WHDLD), and the proposed one obtained the best performance. We also verified the effectiveness of each component of the network through ablation experiments. The code and model will be available at https://github.com/Zhzhyd/DWin-HRFormer.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • AWFLN: An Adaptive Weighted Feature Learning Network for Pansharpening

    • Free pre-print version: Loading...

      Authors: Hangyuan Lu;Yong Yang;Shuying Huang;Xiaolong Chen;Biwei Chi;Aizhu Liu;Wei Tu;
      Pages: 1 - 15
      Abstract: Deep learning (DL)-based pansharpening methods have shown great advantages in extracting spectral–spatial features from multispectral (MS) and panchromatic (PAN) images compared with traditional methods. However, most DL-based methods ignore the local inner connection between the source images and the high-resolution MS (HRMS) image, which cannot fully extract spectral–spatial information and attempt to improve the quality of fusion by increasing the complexity of the network. To solve these problems, a lightweight network based on adaptive weighted feature learning network (AWFLN) is proposed for pansharpening. Specifically, a novel detail extraction model is first built by exploring the local relationship between HRMS and source images, thereby improving the accuracy of details and the interpretability of the network. Guided by this model, we then design a residual multiple receptive-field structure to fully extract spectral–spatial features of source images. In this structure, an adaptive feature learning block based on spectral–spatial interleaving attention is proposed to adaptively learn the weights of features and improve the accuracy of the extracted details. Finally, the pansharpened result is obtained by a detail injection model in AWFLN. Numerous experiments are carried out to validate the effectiveness of the proposed method. Compared to traditional and state-of-the-art methods, AWFLN performs the best both subjectively and objectively, with high efficiency. The code is available at https://github.com/yotick/AWFLN.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Synergistical Attention Model for Semantic Segmentation of Remote
           Sensing Images

    • Free pre-print version: Loading...

      Authors: Xin Li;Feng Xu;Fan Liu;Xin Lyu;Yao Tong;Zhennan Xu;Jun Zhou;
      Pages: 1 - 16
      Abstract: In remotely sensed images, high intraclass variance and interclass similarity are ubiquitous due to complex scenes and objects with multivariate features, making semantic segmentation a challenging task. Deep convolutional neural networks can solve this problem by modeling the context of features and improving their discriminability. However, current learning paradigms model the feature affinity in spatial dimension and channel dimension separately and then fuse them in a sequential or parallel manner, leading to suboptimal performance. In this study, we first analyze this problem practically and summarize it as attention bias that reduces the capability of network in distinguishing weak and discretely distributed objects from wide-range objects with internal connectivity, when modeled only in spatial or channel domain. To jointly model both spatial and channel affinity, we design a synergistic attention module (SAM), which allows for channelwise affinity extraction while preserving spatial details. In addition, we propose a synergistic attention perception neural network (SAPNet) for the semantic segmentation of remote sensing images. The hierarchical-embedded synergistic attention perception module aggregates SAM-refined features and decoded features. As a result, SAPNet enriches inference clues with desired spatial and channel details. Experiments on three benchmark datasets show that SAPNet is competitive in accuracy and adaptability compared with state-of-the-art methods. The experiments also validate the hypothesis of attention bias and the efficiency of SAM.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Semisupervised Change Detection With Feature-Prediction Alignment

    • Free pre-print version: Loading...

      Authors: Xueting Zhang;Xin Huang;Jiayi Li;
      Pages: 1 - 16
      Abstract: Change detection (CD) has received raising attention for its broad application value. However, traditional fully supervised CD methods have a huge demand for pixel-level annotations, which are laborious and even impossible in some few-shot scenarios. Recently, several semisupervised CD (SSCD) methods have been proposed to utilize numerous unlabeled remote sensing image (RSI) pairs, which can largely reduce the annotation dependence. These methods are mainly based on: 1) adversarial learning, whose optimization direction is difficult to control as a black-box method, or 2) feature-consistency learning, which has no explicit physical meaning. To deal with these difficulties, we propose a novel progressive SSCD framework in this article, termed feature-prediction alignment (FPA). FPA can efficiently utilize unlabeled RSI pairs for training by two alignment strategies. First, a class-aware feature alignment (FA) strategy is designed to align the area-level change/no-change feature extracted from different unlabeled RSI pairs (i.e., across regions) with the awareness of their locations, in order to reduce the feature difference within the same classes. Second, a pixelwise prediction alignment (PA) is devised to align the pixel-level change prediction of strongly augmented unlabeled RSI pairs to the pseudo-labels calculated from the corresponding weakly augmented counterparts, in order to reduce the prediction uncertainty of various RSI transformations with physical meaning. Experiments are carried out on four widely used CD benchmarks, including Learning, Vision and Remote Sensing Laboratory (LEVIR-CD), Wuhan University building CD (WHU-CD), CDD, and GZ-CD, and our FPA achieves the state-of-the-art performance. The experimental results demonstrate the superiority of our method in both effectiveness and generalization. Our code is available at https://github.com/zxt9/FPA-SSCD.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multipatch Progressive Pansharpening With Knowledge Distillation

    • Free pre-print version: Loading...

      Authors: Meiqi Gong;Hao Zhang;Han Xu;Xin Tian;Jiayi Ma;
      Pages: 1 - 15
      Abstract: In this article, we propose a novel multipatch and multistage pansharpening method with knowledge distillation, termed PSDNet. Different from the existing pansharpening methods that typically input single-size patches to the network and implement pansharpening in an overall stage, we design multipatch inputs and a multistage network for more accurate and finer learning. First, multipatch inputs allow the network to learn more accurate spatial and spectral information by reducing the number of object types. We employ small patches in the early part to learn accurate local information, as small patches contain fewer object types. Then, the later part exploits large patches to fine-tune it for the overall information. Second, the multistage network is designed to reduce the difficulty of the previous single-step pansharpening and progressively generate elaborate results. In addition, instead of the traditional perceptual loss, which hardly relates to the specific task or the designed network, we introduce distillation loss to reinforce the guidance of the ground truth. Extensive experiments are conducted to demonstrate the superior performance of our proposed PSDNet to the existing state-of-the-art methods. Our code is available at https://github.com/Meiqi-Gong/PSDNet.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multiresolution Analysis Pansharpening Based on Variation Factor for
           Multispectral and Panchromatic Images From Different Times

    • Free pre-print version: Loading...

      Authors: Peng Wang;Hongyu Yao;Bo Huang;Henry Leung;Pengfei Liu;
      Pages: 1 - 17
      Abstract: Most pansharpening methods refer to the fusion of the original low-resolution multispectral (MS) and high-resolution panchromatic (PAN) images acquired simultaneously over the same area. Due to its good robustness, multiresolution analysis (MRA) has become one of the important categories of pansharpening methods. However, when only MS and PAN images acquired at different times can be provided, the fusion results from current MRA methods are often not ideal due to the failure to effectively analyze multitemporal misalignments between MS and PAN images from different times. To solve this issue, MRA pansharpening based on variation factor for MS and PAN images from different times is proposed. The MRA pansharpening based on dual-scale regression model is first established, and the variation factor is then introduced to effectively analyze the multitemporal misalignments by using the alternating direction method of multipliers (ADMM), yielding the final fusion results. Experiments with synthetic and real datasets show that the proposed method exhibits significant performance improvement compared to the traditional pansharpening methods, as well as the state-of-the-art MRA methods. Visual comparisons demonstrate that the variation factor introduces encouraging improvements in the compensation of multitemporal misalignments in ground objects and advances pansharpening applications for MS and PAN images acquired at different times.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Cloud-Target Calibration for Fengyun-3D MERSI-II Solar Reflectance Bands:
           Model Development and Instrument Stability

    • Free pre-print version: Loading...

      Authors: Fukun Wang;Chao Liu;Bin Yao;Xiuqing Hu;Peng Zhang;Byung-Ju Sohn;
      Pages: 1 - 13
      Abstract: Radiative calibration of satellite spectral radiometers is essential for their downstream applications. The Medium Resolution Spectral Imager (MERSI-II) is a key instrument of the Chinese polar orbit Fengyun-3D (FY-3D) satellite. However, its calibration performance has not been sufficiently studied, which limits its broad application. This study revealed the feasibility of a cloud-target method for assessing the MERSI-II calibration performance in solar bands. The top-of-atmosphere (TOA) reflectances for six MERSI-II reflective solar bands (RSBs) were numerically simulated using a rigorous forward radiative transfer method and cloud properties from well-collocated and well-calibrated Moderate Resolution Imaging Spectroradiometer (MODIS) operational cloud products with strict constraints. Only ice cloud targets were examined in the collocation due to their better homogeneity. The excellent agreement between our simulated reflectance and the MODIS reflectance (relative differences (RDs) of over 90% are within a 5% uncertainty range in six bands) validates our models. The simulated results in MERSI-II bands 1–4 showed reasonable agreements with the MERSI-II operational reflectance, i.e., mean RDs < 3%, while the RDs in bands 6 and 7 reaches 12% and 6%, respectively. Our systematic cloud-target-calibration results over three years (2019–2021) indicated clear seasonal calibration biases and signal degradation of the MERSI-II solar bands, and those in the two cloud-absorbing bands, which reached $sim $ 15% and $sim $ 12% (in the three years), respectively. More importantly, we removed these seasonal and degradation biases to improve the cur-ent calibration accuracy to a stable value within 3%. Due to its robust performance, our cloud-target-based calibration method can be applied to future MERSI-II sensors to monitor solar band stability.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • High-Order Semantic Decoupling Network for Remote Sensing Image Semantic
           Segmentation

    • Free pre-print version: Loading...

      Authors: Chengyu Zheng;Jie Nie;Zhaoxin Wang;Ning Song;Jingyu Wang;Zhiqiang Wei;
      Pages: 1 - 15
      Abstract: Low-order features based on convolution kernel are easy to be distorted when encountering dramatic view angle transformation and atmospheric scattering in remote sensing (RS) images. To address this concern, this article first proposes to operate semantic segmentation of RS images based on the high-order information, which can represent the relative relationship of low-order features and is robust and stable when suffering feature distortion. Besides, semantic decouples have recently been well researched and have achieved significant improvement in image understanding. Thus, in this article, a high-order semantic decoupling network (HSDN) is proposed to disentangle features by semantics based on high-order features. Specifically, HSDN first represents each pixel by calculating the pixel-level affinity as a high-order feature and then clusters these pixels into different semantics. Afterward, an attention-like mask generation module is designed for both intra-semantic and inter-semantic groups, leading to three kinds of masks, including the semantic decoupling mask (SDM), which utilizes each high-order cluster centroid as a mask to compact features intracluster and expand different interclusters, so as to improve semantic disentangle performance to a better extent; semantic enhancement mask (SEM), which records pixel-level relative correlation within a class to sufficiently exploit high-order features and could enhance feature robustness; and boundary supplementary mask (BSM), which aims to process borderline pixels to reduce cluster errors. Finally, by applying masks on pixels both within classes and on borderlines, semantic decoupled features are generated and concatenated to realize segmentation. The quantitative and qualitative experiments are conducted on two large-scale fine-resolution RS image datasets to demonstrate the significant performance of adopting high-order representation. Besides, we also implement numerous experiments to va-idate the effectiveness of the proposed semantic decouple framework in dealing with complicated and distortion-prone RS image segmentation tasks.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Detail Injection-Based Spatio-Temporal Fusion for Remote Sensing Images
           With Land Cover Changes

    • Free pre-print version: Loading...

      Authors: Qiang Liu;Xiangchao Meng;Xinghua Li;Feng Shao;
      Pages: 1 - 14
      Abstract: Spatio-temporal fusion can generate time-series images with high spatial resolution, and it is highly desirable in various applications, especially in monitoring fine dynamic changes of surface features on remote sensing images. Currently, most spatio-temporal fusion methods predict the target fine image by employing the auxiliary fine images on neighboring phases; however, they are generally limited in abrupt land cover changes between the target and the neighboring auxiliary images. In this article, we propose a novel detail injection-based spatio-temporal fusion (DISTF) model to alleviate this problem, by exploring the inherent relationship between the spatio-temporal fusion and spatio-spectral fusion. The proposed DISTF consists of three modules: a three-branch detail injection (TDI) module, a fine detail prediction (FDP) module, and a reconstruction module. The interpretable TDI module is inspired by spatio-spectral fusion, aiming to inject the non-changed detail information extracted from the neighboring fine images into the target coarse image, which can preserve the abrupt change information captured in the target coarse image. The FDP module is designed to further integrate the correlated information from the outputs of TDI and refine the spatial-spectral information to boost the fusion accuracy. Finally, the reconstruction module and the hybrid loss function are designed to more effective reconstruct the high-quality target fine image. The qualitative and quantitative experimental results on two datasets with different types of changes demonstrated that the proposed DISTF method achieves richer spatial detail and more accurate prediction than the eight existing methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Sal²RN: A Spatial–Spectral Salient Reinforcement Network for
           Hyperspectral and LiDAR Data Fusion Classification

    • Free pre-print version: Loading...

      Authors: Jiaojiao Li;Yuzhe Liu;Rui Song;Yunsong Li;Kailiang Han;Qian Du;
      Pages: 1 - 14
      Abstract: Hyperspectral image (HSI) and light detection and ranging (LiDAR) data fusion have been widely employed in HSI classification to promote interpreting performance. In the existing deep learning methods based on spatial–spectral features, the features extracted from different layers are treated fairly in the learning process. In reality, features extracted from the continuous layers contribute differentially to the final classification, such as large tracts of woodland and agriculture typically count on shallow contour features, whereas deep semantic spectral features have meaningful constraints for small entities like vehicles. Furthermore, the majority of existing classification algorithms employ a patch input scheme, which has a high probability to introduce pixels of different categories at the boundary. To acquire more accurate classification results, we propose a spatial–spectral saliency reinforcement network (Sal2RN) in this article. In spatial dimension, a novel cross-layer interaction module (CIM) is presented to adaptively alter the significance of features between various layers and integrate these diversified features. Moreover, a customized center spectrum correction module (CSCM) integrates neighborhood information and adaptively modifies the center spectrum to reduce intraclass variance and further improve the classification accuracy of the network. Finally, a statistically based feature weighted combination module is constructed to effectively fuse spatial, spectral, and LiDAR features. Compared with traditional and advanced classification methods, the Sal2RN achieves the state-of-the-art classification performance on three open benchmark datasets.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Implicit Neural Representation Learning for Hyperspectral Image
           Super-Resolution

    • Free pre-print version: Loading...

      Authors: Kaiwei Zhang;Dandan Zhu;Xiongkuo Min;Guangtao Zhai;
      Pages: 1 - 12
      Abstract: Hyperspectral image (HSI) super-resolution (SR) without additional auxiliary image remains a constant challenge due to its high-dimensional spectral patterns, where learning an effective spatial and spectral representation is a fundamental issue. Recently, implicit neural representations (INRs) are making strides as a novel and effective representation, especially in the reconstruction task. Therefore, in this work, we propose a novel HSI reconstruction model based on INR which represents HSI by a continuous function mapping a spatial coordinate to its corresponding spectral radiance values. In particular, as a specific implementation of INR, the parameters of the parametric model are predicted by a hypernetwork that operates on feature extraction using a convolution network. It makes the continuous functions map the spatial coordinates to pixel values in a content-aware manner. Moreover, periodic spatial encoding is deeply integrated with the reconstruction procedure, which makes our model capable of recovering more high-frequency details. To verify the efficacy of our model, we conduct experiments on three HSI datasets (CAVE, NUS, and NTIRE2018). Experimental results show that the proposed model can achieve competitive reconstruction performance in comparison with the state-of-the-art methods. In addition, we provide an ablation study on the effect of individual components of our model. We hope this article could serve as a potent reference for future research.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • MR-Selection: A Meta-Reinforcement Learning Approach for Zero-Shot
           Hyperspectral Band Selection

    • Free pre-print version: Loading...

      Authors: Jie Feng;Gaiqin Bai;Di Li;Xiangrong Zhang;Ronghua Shang;Licheng Jiao;
      Pages: 1 - 20
      Abstract: Band selection is an effective method to deal with the difficulties in image transmission, storage, and processing caused by redundant and noisy bands in hyperspectral images (HSIs). Existing band selection methods usually need to learn a specific model for each HSI dataset, which ignores the inherent correlation and common knowledge among different band selection tasks. Meanwhile, these methods lead to a huge waste of computation. In this article, a novel zero-shot band selection method, called MR-Selection, is proposed for HSI classification. It formalizes zero-shot band selection as a metalearning problem, where advantage actor–critic algorithm-based reinforcement learning (A2C-RL) is designed to extract the metaknowledge in the band selection tasks of various seen hyperspectral datasets through a shared agent. To learn a consistent representation among different tasks, a dynamic structure-aware graph convolutional network is constructed to build a shared agent in A2C-RL. In A2C-RL, the state is tailored in a feasible way and easy to adapt to various tasks. Meanwhile, the reward is defined according to an efficient evaluation network, which can evaluate each state effectively without any fine-tuning. Furthermore, a two-stage optimization strategy is designed to coordinate optimization directions of a shared agent from different tasks effectively. Once the shared agent is optimized, it can be directly applied to unseen HSI band selection tasks without any available samples. Experimental results demonstrate the effectiveness and efficiency of the MR-Selection on the band selection of unseen HSI datasets.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Hyperspectral Sparse Unmixing via Nonconvex Shrinkage Penalties

    • Free pre-print version: Loading...

      Authors: Longfei Ren;Danfeng Hong;Lianru Gao;Xu Sun;Min Huang;Jocelyn Chanussot;
      Pages: 1 - 15
      Abstract: Hyperspectral sparse unmixing aims at finding the optimal subset of spectral signatures in the given spectral library and estimating their proportions in each pixel. Recently, simultaneously sparse and low-rank representations (SSLRRs) have been widely used in the hyperspectral sparse unmixing task. This article developed a new unified framework to approximate the SSLRR-based unmixing model. The heart of the proposed framework is to design the new nonconvex penalties for efficient minimization by the means of two families of thresholding mappings, including the firm thresholding mapping and generalized shrinkage mapping. Both mappings can be regarded as generalizations of both soft thresholding and hard thresholding. Unlike previous approaches with explicit penalties, the proposed framework does not require explicit forms of penalties but only relevant thresholding mappings. Furthermore, an alternating direction method of multipliers (ADMM) was designed to solve the resulting optimization problem. Experiments conducted on the synthetic data and real data demonstrate the superiority of the proposed framework in improving the unmixing performance with respect to state-of-the-art methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Hyperspectral Remote Sensing Image Synthesis Based on Implicit Neural
           Spectral Mixing Models

    • Free pre-print version: Loading...

      Authors: Liqin Liu;Zhengxia Zou;Zhenwei Shi;
      Pages: 1 - 14
      Abstract: Hyperspectral image (HSI) synthesis, as an emerging research topic, is of great value in overcoming sensor limitations and achieving low-cost acquisition of high-resolution remote sensing HSIs. However, the linear spectral mixing model used in recent studies oversimplifies the real-world hyperspectral imaging process, making it difficult to effectively model the imaging noise and multiple reflections of the object spectrum. As a prerequisite for hyperspectral data synthesis, accurate modeling of nonlinear spectral mixtures has long been a challenge. Considering the above difficulties, we propose a novel method for modeling nonlinear spectral mixtures based on implicit neural representations (INRs) in this article. The proposed method learns from INR and adaptively implements different mixture models for each pixel according to their spectral signature and surrounding environment. Based on the above neural mixing model, we also propose a new method for HSI synthesis. Given an RGB image as input, our method can generate an accurate and physically meaningful HSI. As a set of by-products, our method can also generate subpixel-level spectral abundance as well as the solar atmosphere signature. The whole framework is trained end-to-end in a self-supervised manner. We constructed a new dataset for HSI synthesis based on a wide range of Airborne Visible Infrared Imaging Spectrometer (AVIRIS) data. Our method achieves a mean peak signal-to-noise ratio (MPSNR) of 52.36 dB and outperforms other state-of-the-art hyperspectral synthesis methods. Finally, our method shows great benefits to downstream data-driven applications. With the HSIs and abundance directly generated from low-cost RGB images, the proposed method improves the accuracy of HSI classification tasks by a large margin, particularly for those with limited training samples.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Bayesian Meta-Learning-Based Method for Few-Shot Hyperspectral Image
           Classification

    • Free pre-print version: Loading...

      Authors: Jing Zhang;Liqin Liu;Rui Zhao;Zhenwei Shi;
      Pages: 1 - 13
      Abstract: Few-shot learning provides a new way to solve the problem of insufficient training samples in hyperspectral classification. It can implement reliable classification under several training samples by learning meta-knowledge from similar tasks. However, most existing works perform frequency statistics, which may suffer from the prevalent uncertainty in point estimates (PEs) with limited training samples. To overcome this problem, we reconsider the hyperspectral image few-shot classification (HSI-FSC) task as a hierarchical probabilistic inference from a Bayesian view and provide a careful process of meta-learning probabilistic inference. We introduce a prototype vector for each class as latent variables and adopt distribution estimates (DEs) for them to obtain their posterior distribution. The posterior of the prototype vectors is maximized by updating the parameters in the model via the prior distribution of HSI and labeled samples. The features of the query samples are matched with prototype vectors drawn from the posterior; thus, a posterior predictive distribution over the labels of query samples can be inferred via an amortized Bayesian variational inference approach. Experimental results on four datasets demonstrate the effectiveness of our method. Especially given only three to five labeled samples, the method achieves noticeable upgrades of overall accuracy (OA) against competitive methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Joint Classification of Hyperspectral and LiDAR Data Using a Hierarchical
           CNN and Transformer

    • Free pre-print version: Loading...

      Authors: Guangrui Zhao;Qiaolin Ye;Le Sun;Zebin Wu;Chengsheng Pan;Byeungwoo Jeon;
      Pages: 1 - 16
      Abstract: The joint use of multisource remote-sensing (RS) data for Earth observation missions has drawn much attention. Although the fusion of several data sources can improve the accuracy of land-cover identification, many technical obstacles, such as disparate data structures, irrelevant physical characteristics, and a lack of training data, exist. In this article, a novel dual-branch method, consisting of a hierarchical convolutional neural network (CNN) and a transformer network, is proposed for fusing multisource heterogeneous information and improving joint classification performance. First, by combining the CNN with a transformer, the proposed dual-branch network can significantly capture and learn spectral–spatial features from hyperspectral image (HSI) data and elevation features from light detection and ranging (LiDAR) data. Then, to fuse these two sets of data features, a cross-token attention (CTA) fusion encoder is designed in a specialty. The well-designed deep hierarchical architecture takes full advantage of the powerful spatial context information extraction ability of the CNN and the strong long-range dependency modeling ability of the transformer network based on the self-attention (SA) mechanism. Four standard datasets are used in experiments to verify the effectiveness of the approach. The experimental results reveal that the proposed framework can perform noticeably better than state-of-the-art methods. The source code of the proposed method will be available publicly at https://github.com/zgr6010/Fusion_HCT.git.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Edge-Guided Hyperspectral Image Compression With Interactive Dual
           Attention

    • Free pre-print version: Loading...

      Authors: Yuanyuan Guo;Yulong Tao;Yanwen Chong;Shaoming Pan;Miao Liu;
      Pages: 1 - 17
      Abstract: Compressing hyperspectral images (HSIs) into compact representations under the premise of ensuring high-quality reconstruction is an essential task in HSI processing. However, existing compression methods usually encode images by smoothing due to the low-frequency information occupying a prominent component in most images. Consequently, these methods fail to capture sufficient structural information, especially in low bit rates, often causing inferior reconstruction. To address this problem, we propose here an edge-guided hyperspectral compression network, called CENet, to realize high-quality reconstruction. To enhance the structural latent representation ability, the CENet model incorporates an edge extractor neural network into the compression architecture to guide compression optimization by the edge-guided loss. We propose an interactive dual attention module to selectively learn edge features, obtain the most effective edge structure, and avoid additional edge information redundancy at the same time. In the proposed CENet, the edge-guided loss and interactive dual attention module are combined to enhance the comprehensive structure of the latent representation. Concretely, interactive dual attention makes the edge extraction network focus only on moderate boundaries rather than on all edges, which enables savings on the bit rate cost and helps achieve a strong structural representation. As a result, the reconstruction quality is significantly improved. The extensive experiments on seven HSI datasets verify that our model can effectively raise the rate–distortion performance for HSIs of any type or resolution (e.g., yielding an average peak signal-to-noise ratio (PSNR) of 30.59 dB at 0.2382 bpppb, which exceeds the baseline for Chikusei by 10.99%).
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Novel Hyperspectral Image Classification Model Using Bole Convolution
           With Three-Direction Attention Mechanism: Small Sample and Unbalanced
           Learning

    • Free pre-print version: Loading...

      Authors: Weiwei Cai;Xin Ning;Guoxiong Zhou;Xiao Bai;Yizhang Jiang;Wei Li;Pengjiang Qian;
      Pages: 1 - 17
      Abstract: Currently, the use of rich spectral and spatial information of hyperspectral images (HSIs) to classify ground objects is a research hotspot. However, the classification ability of existing models is significantly affected by its high data dimensionality and massive information redundancy. Therefore, we focus on the elimination of redundant information and the mining of promising features and propose a novel Bole convolution (BC) neural network with a tandem three-direction attention (TDA) mechanism (BTA-Net) for the classification of HSI. A new BC is proposed for the first time in this algorithm, whose core idea is to enhance effective features and eliminate redundant features through feature punishment and reward strategies. Considering that traditional attention mechanisms often assign weights in a one-direction manner, leading to a loss of the relationship between the spectra, a novel three-direction (horizontal, vertical, and spatial directions) attention mechanism is proposed, and an addition strategy and a maximization strategy are used to jointly assign weights to improve the context sensitivity of spatial–spectral features. In addition, we also designed a tandem TDA mechanism module and combined it with a multiscale BC output to improve classification accuracy and stability even when training samples are small and unbalanced. We conducted scene classification experiments on four commonly used hyperspectral datasets to demonstrate the superiority of the proposed model. The proposed algorithm achieves competitive performance on small samples and unbalanced data, according to the results of comparison and ablation experiments. The source code for BTA-Net can be found at https://github.com/vivitsai/BTA-Net.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • t-Linear Tensor Subspace Learning for Robust Feature Extraction of
           Hyperspectral Images

    • Free pre-print version: Loading...

      Authors: Yang-Jun Deng;Heng-Chao Li;Si-Qiao Tan;Junhui Hou;Qian Du;Antonio Plaza;
      Pages: 1 - 15
      Abstract: Subspace learning has been widely applied for feature extraction of hyperspectral images (HSIs) and achieved great success. However, the current methods still leave two problems that need to be further investigated. First, those methods mainly focus on finding one or multiple projection matrices for mapping the high-dimensional data into a low-dimensional subspace, which can only capture the information from each direction of high-order hyperspectral data separately. Second, the performance of feature extraction is barely satisfactory when the hyperspectral data is severely corrupted by noise. To address these issues, this article presents a t-linear tensor subspace learning (tLTSL) model for robust feature extraction of HSIs based on t-product projection. In the model, t-product projection is a new defined tensor transformation way similar to linear transformation in vector space, which can maximally capture the intrinsic structure of tensor data. The integrated tensor low-rank and sparse decomposition can effectively remove the noise corruption and the learned t-product projection can directly transform the high-order hyperspectral data into a subspace with information from all modes comprehensively considered. Moreover, a proposition related to tensor rank is proofed for interpreting the meaning of the tLTSL model. Extensive experiments are conducted on two different kinds of noise (i.e., simulated and real noise) corrupted HSI data, which validate the effectiveness of tLTSL.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Graph Meta Transfer Network for Heterogeneous Few-Shot Hyperspectral Image
           Classification

    • Free pre-print version: Loading...

      Authors: Haoyu Wang;Xuesong Wang;Yuhu Cheng;
      Pages: 1 - 12
      Abstract: Since obtaining labeled hyperspectral images (HSIs) is difficult and time-consuming, the shortage of training samples has always been a challenge for HSI classification. In practical applications, only a few labeled samples are available in the task domain (target domain), while sufficient labeled samples are available in another domain (source domain). At the same time, these two domains are heterogeneous and contain different categories. This scenario makes it difficult to effectively transfer knowledge from the source domain to the target domain. To address this challenge, we propose a novel heterogeneous few-shot learning (FSL) method, namely graph meta transfer network (GMTN). Specifically, the graph sample and aggregate network (GraphSAGE) and meta-learning, which are both inductive learning, are integrated into a unified framework. In this way, the aggregation function is generalized from abundant few-shot tasks for feature extraction on the source and target domains. The spatial importance strategy (SIS) is designed to guide the feature propagation and alleviate the information interference caused by different categories. The neighborhood receptive field spectral attention (RFSA) mechanism is proposed to model the importance of spectral band using the information of the neighborhood pixels, which enables GMTN to pay more attention to bands with discriminative features in both domains. In addition, the node spatial information reset method is proposed to augment samples based on the spatial position relationship of nodes. Furthermore, to alleviate the domain shift in heterogeneous scenarios, the conditional domain adversarial strategy is used to achieve effective meta-knowledge transfer. Experiments show that GMTN outperforms the compared state-of-the-art methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Morphological Transformation and Spatial-Logical Aggregation for Tree
           Species Classification Using Hyperspectral Imagery

    • Free pre-print version: Loading...

      Authors: Mengmeng Zhang;Wei Li;Xudong Zhao;Huan Liu;Ran Tao;Qian Du;
      Pages: 1 - 12
      Abstract: Hyperspectral image (HSI) consists of abundant spectral and spatial characteristics, which contribute to a more accurate identification of materials and land covers. However, most existing methods of hyperspectral image analysis primarily focus on spectral knowledge or coarse-grained spatial information while neglecting the fine-grained morphological structures. In the classification task of complex objects, spatial morphological differences can help to search for the boundary of fine-grained classes, e.g., forestry tree species. Focusing on subtle traits extraction, a spatial-logical aggregation network (SLA-NET) is proposed with morphological transformation for tree species classification. The morphological operators are effectively embedded with the trainable structuring elements, which contributes to distinctive morphological representations. We evaluate the classification performance of the proposed method on two tree species datasets, and the results demonstrate that the proposed SLA-NET significantly outperforms the other state-of-the-art classifiers.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Language-Aware Domain Generalization Network for Cross-Scene Hyperspectral
           Image Classification

    • Free pre-print version: Loading...

      Authors: Yuxiang Zhang;Mengmeng Zhang;Wei Li;Shuai Wang;Ran Tao;
      Pages: 1 - 12
      Abstract: Text information including extensive prior knowledge about land cover classes has been ignored in hyperspectral image (HSI) classification tasks. It is necessary to explore the effectiveness of linguistic mode in assisting HSI classification. In addition, the large-scale pretraining image–text foundation models have demonstrated great performance in a variety of downstream applications, including zero-shot transfer. However, most domain generalization methods have never addressed mining linguistic modal knowledge to improve the generalization performance of model. To compensate for the inadequacies listed above, a language-aware domain generalization network (LDGnet) is proposed to learn cross-domain-invariant representation from cross-domain shared prior knowledge. The proposed method only trains on the source domain (SD) and then transfers the model to the target domain (TD). The dual-stream architecture including the image encoder and text encoder is used to extract visual and linguistic features, in which coarse-grained and fine-grained text representations are designed to extract two levels of linguistic features. Furthermore, linguistic features are used as cross-domain shared semantic space, and visual–linguistic alignment is completed by supervised contrastive learning in semantic space. Extensive experiments on three datasets demonstrate the superiority of the proposed method when compared with the state-of-the-art techniques. The codes will be available from the website: https://github.com/YuxiangZhang-BIT/IEEE_TGRS_LDGnet.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Lightweight Multiresolution Feature Fusion Network for Spectral
           Super-Resolution

    • Free pre-print version: Loading...

      Authors: Shaohui Mei;Ge Zhang;Nan Wang;Bo Wu;Mingyang Ma;Yifan Zhang;Yan Feng;
      Pages: 1 - 14
      Abstract: Spectral super-resolution (SR), which reconstructs high spatial-resolution hyperspectral images (HSIs) from RGB inputs, has been demonstrated to be one of the effective computational imaging techniques to acquire HSIs. Though deep neural networks have shown their superiority in such a complex mapping problem, existing networks generally involve a very complex structure with huge amounts of parameters, resulting in giant memory occupation. In this article, a lightweight multiresolution feature fusion network (MRFN) is proposed, which adopts a multiresolution feature extraction and fusion framework to fully explore RGB inputs in different scales of resolution. Specifically, a lightweight feature extraction module (LFEM), which adopts cheap convolution and attention mechanisms, is constructed to explore different scales of features under a lightweight structure. Moreover, a hybrid loss function is proposed by encountering not only pixel-value level reconstruction error but also spectral continuity and fidelity. Experiments over three benchmark datasets, i.e., CAVE, Interdisciplinary Computational Vision Laboratory (ICVL), and NTIRE2022 datasets, have demonstrated that the proposed MRFN can reconstruct HSIs from RGB inputs in higher quality with fewer parameters and computational floating-point operations (FLOPs) compared with several state-of-the-art networks.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Abundance Matrix Correlation Analysis Network Based on Hierarchical
           Multihead Self-Cross-Hybrid Attention for Hyperspectral Change Detection

    • Free pre-print version: Loading...

      Authors: Wenqian Dong;Jingyu Zhao;Jiahui Qu;Song Xiao;Nan Li;Shaoxiong Hou;Yunsong Li;
      Pages: 1 - 13
      Abstract: Hyperspectral image (HSI) change detection is a technique for detecting the changes between the multitemporal HSIs of the same scene. Many existing change detection methods have achieved good results, but there still exist problems as follows: 1) mixed pixels exist in HSI due to the low spatial resolution of hyperspectral sensor and other external interference and 2) many existing deep learning-based networks cannot make full use of the correlation difference information between the bitemporal images. These problems are not conducive to further improving the accuracy of change detection. In this article, we propose an abundance matrix correlation analysis network based on hierarchical multihead self-cross-hybrid attention (AMCAN-HMSchA) for HSI change detection, which hierarchically highlights the correlation difference information at the subpixel level to detect the subtle changes. The endmember sharing-based abundance matrix learning module (AMLM) maps the changed information between bitemporal HSIs to the corresponding abundance matrices. The hierarchical MSchA extracts the enhanced difference features by constantly comparing the self-correlation with cross correlation between the abundance matrices of the HSIs. Then, the difference features are concatenated and fed into the fully connected layers to obtain the change map. Experiments on three widely used datasets show that the proposed method has superior performance compared with other state-of-the-art methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multiscale and Cross-Level Attention Learning for Hyperspectral Image
           Classification

    • Free pre-print version: Loading...

      Authors: Fulin Xu;Ge Zhang;Chao Song;Hui Wang;Shaohui Mei;
      Pages: 1 - 15
      Abstract: Transformer-based networks, which can well model the global characteristics of inputted data using the attention mechanism, have been widely applied to hyperspectral image (HSI) classification and achieved promising results. However, the existing networks fail to explore complex local land cover structures in different scales of shapes in hyperspectral remote sensing images. Therefore, a novel network named multiscale and cross-level attention learning (MCAL) network is proposed to fully explore both the global and local multiscale features of pixels for classification. To encounter local spatial context of pixels in the transformer, a multiscale feature extraction (MSFE) module is constructed and implemented into the transformer-based networks. Moreover, a cross-level feature fusion (CLFF) module is proposed to adaptively fuse features from the hierarchical structure of MSFEs using the attention mechanism. Finally, the spectral attention module (SAM) is implemented prior to the hierarchical structure of MSFEs, by which both the spatial context and spectral information are jointly emphasized for hyperspectral classification. Experiments over several benchmark datasets demonstrate that the proposed MCAL obviously outperforms both the convolutional neural network (CNN)-based and transformer-based state-of-the-art networks for hyperspectral classification.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Orthogonal Subspace Unmixing to Address Spectral Variability for
           Hyperspectral Image

    • Free pre-print version: Loading...

      Authors: Longfei Ren;Danfeng Hong;Lianru Gao;Xu Sun;Min Huang;Jocelyn Chanussot;
      Pages: 1 - 13
      Abstract: Hyperspectral unmixing aims at estimating pure spectral signatures and their proportions in each pixel. In practice, the atmospheric effects, intrinsic variation of the spectral signatures of the materials, illumination, and topographic changes cause what is known as spectral variability resulting in significant estimation errors being propagated throughout the unmixing task. To this end, we developed a new method, called the orthogonal subspace unmixing (OSU), to address spectral variability by utilizing the orthogonal subspace projection. The proposed OSU method jointly performs orthogonal subspace learning and the unmixing process to find a more suitable subspace for unmixing. The orthogonal subspace projection encourages the representation held in the subspace to be more distinct from each other to remove the complex spectral variability in the subspace. Furthermore, an alternating minimization (AM) was designed to solve the resulting optimization problem. An efficient and convergent symmetric Gauss–Seidel alternating direction method of multipliers (sGS-ADMM), essentially a special case of the semiproximal alternating direction method of multipliers (SPADMM), was developed to solve the subproblem. Experiments conducted on one synthetic data and two real data demonstrate the effectiveness and superiority of the proposed framework in mitigating the effects of spectral variability with respect to classical linear unmixing methods or variability accounting approaches.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Nearest Neighbor-Based Contrastive Learning for Hyperspectral and LiDAR
           Data Classification

    • Free pre-print version: Loading...

      Authors: Meng Wang;Feng Gao;Junyu Dong;Heng-Chao Li;Qian Du;
      Pages: 1 - 16
      Abstract: The joint hyperspectral image (HSI) and light detection and ranging (LiDAR) data classification aims to interpret ground objects at more detailed and precise level. Although deep learning methods have shown remarkable success in the multisource data classification task, self-supervised learning has rarely been explored. It is commonly nontrivial to build a robust self-supervised learning model for multisource data classification, due to the fact that the semantic similarities of neighborhood regions are not exploited in the existing contrastive learning framework. Furthermore, the heterogeneous gap induced by the inconsistent distribution of multisource data impedes the classification performance. To overcome these disadvantages, we propose a nearest neighbor-based contrastive learning network (NNCNet), which takes full advantage of large amounts of unlabeled data to learn discriminative feature representations. Specifically, we propose a nearest neighbor-based data augmentation scheme to use enhanced semantic relationships among nearby regions. The intermodal semantic alignments can be captured more accurately. In addition, we design a bilinear attention module to exploit the second-order and even high-order feature interactions between the HSI and LiDAR data. Extensive experiments on four public datasets demonstrate the superiority of our NNCNet over state-of-the-art methods. The source codes are available at https://github.com/summitgao/NNCNet.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Lightweighted Hyperspectral Image Classification Network by Progressive
           Bi-Quantization

    • Free pre-print version: Loading...

      Authors: Wei Wei;Chongxing Song;Lei Zhang;Yanning Zhang;
      Pages: 1 - 14
      Abstract: Convolutional neural network (CNN) has shown its powerful ability for hyperspectral image (HSI) classification, which however, is difficult to deploy on resource-limited or low-latency platforms due to its parameter and computation redundancy. Though binary neural network (BNN) has attracted attention for its extreme compressing and speeding up ability by binarizing both weights and activations, it has rarely been explored for HSI classification. In this study, we elaborately design a BNN with good performance for HSI classification task. Specifically, an adaptive gradient scale module is proposed to flexibly modify the gradient during training stage to better optimize the BNN and does not add any extra computation for inference. Furthermore, a curriculum learning-based progressive binarization strategy is utilized to improve the performance. Compared with the existing BNN works, our method can increase the HSI classification accuracy by a large margin while maintaining the compressing ratio. Abundant experiments on three datasets demonstrate the effectiveness of the proposed method.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multiview Spatial–Spectral Two-Stream Network for Hyperspectral
           Image Unmixing

    • Free pre-print version: Loading...

      Authors: Lin Qi;Zhenwei Chen;Feng Gao;Junyu Dong;Xinbo Gao;Qian Du;
      Pages: 1 - 16
      Abstract: Linear spectral unmixing is an important technique in the analysis of mixed pixels in hyperspectral images. In recent years, deep learning-based methods have been garnering increasing attention in hyperspectral unmixing; especially, unsupervised autoencoder (AE) networks that have achieved excellent unmixing performance are a recent trend. While most approaches use spatial information, it is well known that hyperspectral data are characterized by a large number of narrow spectral bands. In order to take full advantage of the hyperspectral bands in unmixing and the spatial information, in this article, we explore multiview spectral and spatial information in an AE-based unmixing framework. We introduce multiview spectral information through spectral partitioning and propose a multiview spatial–spectral two-stream network, called MSSS-Net, which simultaneously learns a spatial stream network and a multiview spectral stream network in an end-to-end fashion for more efficient unmixing. The MSSS-Net is a two-stream deep unmixing network sharing a decoder, where its two AE networks employ recurrent neural networks (RNNs) to collaboratively utilize multiview spectral and spatial information. The spatial stream network branch extracts the spatial features of pixels and its neighbors, while the multiview spectral stream network branch exploits the multiview spectral bands of a pixel. Meanwhile, we design a cascaded bidirectional and unidirectional RNNs’ encoder structure for multiview spatial–spectral information to learn more discriminative deep patch-pixel features. Extensive ablation studies and experiments on both synthetic and real datasets demonstrate the superiority of the MSSS-Net over state-of-the-art unmixing methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Cross-Track Illumination Correction for Hyperspectral Pushbroom Sensor
           Images Using Low-Rank and Sparse Representations

    • Free pre-print version: Loading...

      Authors: Lina Zhuang;Michael K. Ng;Yao Liu;
      Pages: 1 - 17
      Abstract: A hyperspectral pushbroom sensor scans objects line-by-line using a detector array, and a cross-track illumination error (CTIE) exists in the imagery acquired in this way. When the illumination of the individual cells of the detector is not aligned well, or if some of the cells are degraded or old, the acquired images will exhibit nonuniform illumination in the cross-track direction. As additive Gaussian noise is found widely in hyperspectral images (HSIs), we develop a unified mathematical model that describes the image formation process corrupted by the CTIE and additive Gaussian noise. The CTIE produced by line-by-line scanning is replicated and modeled as an offset term with the equivalent values in the direction of flight. The main contribution of this study is the development of a hyperspectral image cross-track illumination correction (HyCIC) method, which corrects the cross-track illumination using column (along-track) mean compensation with total variation and sparsity regularizations, and attenuates the Gaussian noise by using a form of low-rank constraint. The effectiveness of the proposed method is illustrated using semireal data and real HSIs. The performance of the proposed HyCIC is found to be better than other existing methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Spectral–Spatial Prototype Learning-Based Nearest Neighbor Classifier
           for Hyperspectral Images

    • Free pre-print version: Loading...

      Authors: Dan Li;Yuan Shen;Fangqiang Kong;Jiahang Liu;Qiang Wang;
      Pages: 1 - 15
      Abstract: Due to the Hughes phenomenon, hyperspectral image (HSI) classification under small sample size situation is still a key challenging problem. To alleviate this issue, we propose a novel spectral–spatial prototype learning-based nearest neighbor classifier (SSPLNN) for HSI in this article. The local spectral–spatial neighbor set is first constructed for each sample based on both spectral similarity and spatial structural context to accurately explore the local spectral–spatial information. Then, a spectral–spatial prototype learning model is designed to learn a set of spectral–spatial prototypes, which can optimally utilize both the similarity and variance of samples within each spectral–spatial set and excavate the unseen spectral–spatial variations. The learned spectral–spatial prototypes offer more complementary information to improve the classification accuracy remarkably under small sample size situation. In addition, a linear discriminative projection is simultaneously learned to make each test local spectral–spatial set to be optimally classified to the same class with its nearest neighbor (NN) spectral–spatial prototype set in the projected target subspace. Finally, the NN classifier based on measuring the minimum geometric distance between the projected test spectral–spatial set and the projected spectral–spatial prototype sets is employed to determine the label. Experimental results demonstrate that the proposed SSPLNN method outperforms several well-known classification methods by a large margin on three widely analyzed HSI datasets.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Hyperspectral Image Denoising: Reconciling Sparse and Low-Tensor-Ring-Rank
           Priors in the Transformed Domain

    • Free pre-print version: Loading...

      Authors: Hao Zhang;Ting-Zhu Huang;Xi-Le Zhao;Wei He;Jae Kyu Choi;Yu-Bang Zheng;
      Pages: 1 - 13
      Abstract: Recently, the transform-based tensor nuclear norm (TNN) framework has yielded promising results for hyperspectral image (HSI) denoising as compared with previous original-domain tensor-based models. However, the TNN framework only exploits the low-rankness of each band of HSIs (tensors) under a single spectral transform. The correlation between all bands under the transform (i.e., the global low-rankness of the transformed tensor) and the sparsity of the transformed HSI, which are beneficial for HSI denoising, is usually neglected in the TNN framework. In this article, we propose to reconcile sparse and low-tensor-ring (TR)-rank priors in the learned transformed domain (called T-RSTR model) for HSI denoising. In T-RSTR, the transform-based low-TR-rank and sparse regularizers are designed to characterize the global low-rankness and sparsity of the transformed tensors, respectively, and then the transform-based low-TR-rank and sparse regularizers are organically integrated and benefit from each other for substantially boosting denoising performance. To tackle the T-RSTR model, we elaborately design a proximal alternating minimization-based algorithm with the theoretical convergence. Extensive numerical results demonstrate that T-RSTR is superior to the competing methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Cross-Domain Few-Shot Hyperspectral Image Classification With Class-Wise
           Attention

    • Free pre-print version: Loading...

      Authors: Wenzhen Wang;Fang Liu;Jia Liu;Liang Xiao;
      Pages: 1 - 18
      Abstract: Few-shot learning (FSL) is an effective method to solve the problem of hyperspectral image (HSI) classification with few labeled samples. It learns transferable knowledge from sufficient labeled auxiliary data to classify unseen classes with limited labeled samples for training. However, the distribution difference between auxiliary data and unseen classes results in the learned transferable knowledge not being well applied to the new task. Therefore, a class-wise attentive cross-domain FSL (CA-CFSL) framework is proposed in this article, in which a feature extractor is learned to extract data features with discriminability and domain invariance. The class-wise attention metric module (CAMM) introduces class-wise attention on the FSL framework to learn more discriminative features, which improves the interclass decision boundaries. Furthermore, an asymmetric domain adversarial module (ADAM) is designed to enhance the ability of extracting domain-invariant representations, which combines asymmetric adversarial training with embedded domain-specific information. Experimental results on four public HSI datasets demonstrate that the proposed method outperforms the existing methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Realistic Mixing Miniature Scene Hyperspectral Unmixing: From Benchmark
           Datasets to Autonomous Unmixing

    • Free pre-print version: Loading...

      Authors: Chunyang Cui;Yanfei Zhong;Xinyu Wang;Liangpei Zhang;
      Pages: 1 - 15
      Abstract: Mixed pixels that contain more than one material type are common in mid/low spatial resolution remote sensing imagery. Hyperspectral unmixing is aimed at decomposing the mixed pixels into endmembers and abundances. However, there are few datasets that are suitable for quantitatively evaluating unmixing accuracies, and the ground-truth abundances of the existing datasets are often generated in an approximate way. To address the lack of real unmixing datasets for quantitative evaluation, we built the realistic mixing miniature scenes (RMMS) dataset, which can be used to quantitatively evaluate the unmixing accuracy of different algorithms. The RMMS dataset consists of a simple mixture scene with homogeneous flat materials and a complex mixture scene with 3-D structural features. The features of the RMMS dataset also take point, line, and polygon characteristics into consideration, and the spectral similarity of the materials increases the challenge of the spectral unmixing. In the RMMS dataset, due to the multiscale observation characteristics of the spatiotemporal scanning modality, it can avoid the registration error between RGB and hyperspectral data, and it can ensure that the endmembers are pure pixels. Most of the autonomous hyperspectral unmixing algorithms focus on solving some of the unmixing problems and have difficulty achieving fully autonomous hyperspectral unmixing (FAHU). In this article, to overcome this shortcoming, a fully autonomous hyperspectral unmixing method called FAHU is proposed to take advantage of the spatial information. Some of the state-of-the-art autonomous hyperspectral unmixing algorithms are used to evaluate the performance with the RMMS dataset, and the experimental results show the advantages and disadvantages of the different autonomous unmixing algorithms.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Hyperspectral Image Instance Segmentation Using Spectral–Spatial
           Feature Pyramid Network

    • Free pre-print version: Loading...

      Authors: Leyuan Fang;Yifan Jiang;Yinglong Yan;Jun Yue;Yue Deng;
      Pages: 1 - 13
      Abstract: In recent years, hyperspectral image (HSI) classification and detection techniques based on deep learning have been widely applied to various aspects, such as environmental monitoring, urban planning, and energy surveys. As an important image content analysis method, instance segmentation can provide important support for the extraction of ground object information and monomeric application of HSI. This article introduces instance segmentation into HSI interpretation for the first time. In this article, we create the hyperspectral instance segmentation dataset (HS-ISD), which contains a total of 56 images, each with a size of $298times301$ and a number of channels of 48. More than 1000 architectural examples are annotated to apply to the research of HSI instance segmentation. In addition, considering that HSI contains rich spectral and spatial information, and the traditional instance segmentation network model cannot well utilize both types of information effectively, we propose the spectral–spatial feature pyramid network (Spectral–Spatial FPN). The Spectral–Spatial FPN can integrate multiscale spectral information and multiscale spatial information in the feature extraction stage through attention mechanism and bidirectional feature pyramid structure, so as to better improve the performance of the network model by spectral information and spatial information and realize the end-to-end instance segmentation of HSI. The experimental results conducted on the HS-ISD show that the proposed Spectral–Spatial FPN can achieve state-of-the-art results.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Multiscale Diff-Changed Feature Fusion Network for Hyperspectral Image
           Change Detection

    • Free pre-print version: Loading...

      Authors: Fulin Luo;Tianyuan Zhou;Jiamin Liu;Tan Guo;Xiuwen Gong;Jinchang Ren;
      Pages: 1 - 13
      Abstract: For hyperspectral image (HSI) change detection (CD), multiscale features are usually used to construct the detection models. However, the existing studies only consider the multiscale features containing changed and unchanged components, which is difficult to represent the subtle changes between bitemporal HSIs in each scale. To address this problem, we propose a multiscale diff-changed feature fusion network (MSDFFN) for HSI CD, which improves the ability of feature representation by learning the refined change components between bitemporal HSIs under different scales. In this network, a temporal feature encoder–decoder subnetwork, which combines a reduced inception (RI) module and a cross-layer attention module to highlight the significant features, is designed to extract the temporal features of HSIs. A bidirectional diff-changed feature representation (BDFR) module is proposed to learn the fine changed features of bitemporal HSIs at various scales to enhance the discriminative performance of the subtle change. A multiscale attention fusion (MSAF) module is developed to adaptively fuse the changed features of various scales. The proposed method can not only discover the subtle change in bitemporal HSIs but also improve the discriminating power for HSI CD. Experimental results on three HSI datasets show that MSDFFN outperforms a few state-of-the-art methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Progressive Token Reduction and Compensation for Hyperspectral Image
           Representation

    • Free pre-print version: Loading...

      Authors: Chen Ma;Junjun Jiang;Huayi Li;Wenxue Cui;Guoyuan Li;
      Pages: 1 - 14
      Abstract: Hyperspectral images (HSIs) have been widely used in Earth observation because they contain continuous and detailed spectral information which is beneficial for the fine-grained diagnosis of the land cover. In the past few years, convolutional neural network (CNN)-based methods show limitations in modeling spectral-wise long-range dependences. Recently, transformer-based deep learning methods are proposed and have shown superiority in modeling the continuous representation of the spectral signatures because the self-attention (SA) mechanism has a global receptive field. Due to the special tokenization of the transformer-based methods, the redundant tokens contained in spectral embeddings are always involved in SA operation. Redundant tokens do not positively contribute to classification. Specifically, the overlapped group-wise tokenization approach may aggravate the Hughes phenomenon and impose additional computations. To address this issue, a lightweight spatial–spectral pyramid transformer (SSPT) framework is proposed to efficiently extract the spatial–spectral features of HSI by progressively reducing redundant tokens in an end-to-end manner. In particular, a token reduction (TR) method is proposed to decide which tokens will be involved by computing and comparing token attentiveness between spectral embeddings and the class token. In addition, for those tokens that are defined as redundant information, a token compensation mechanism is proposed to automatically extract supplementary information for classification. Extensive experiments on three standard datasets quantitatively show the superiority of our methods, and the ablation experiments qualitatively prove our hypothesis about the feature distribution in transformer architecture.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Adaptive Hypergraph Regularized Multilayer Sparse Tensor Factorization for
           Hyperspectral Unmixing

    • Free pre-print version: Loading...

      Authors: Pan Zheng;Hongjun Su;Hongliang Lu;Qian Du;
      Pages: 1 - 18
      Abstract: Hyperspectral unmixing with tensor models has received great attention in recent years. A tensor-based decomposition method can effectively represent the structural feature of hyperspectral images; however, the obtained results may be physically uninterpretable. To overcome this limitation, a novel adaptive hypergraph regularized multilayer sparse tensor factorization (AHGMLSTF) algorithm is proposed. First, a modified hypergraph is incorporated into tensor factorization, and the modified hypergraph uses spectral angle distance (SAD) instead of Euclidean distance to construct hyperedges to better represent the joint spatial and spectral information. Then, the hypergraph is constructed adaptively by hyperedges of $k$ neighborhoods. Second, the concept of multilayer decomposition is introduced to explore the hierarchical features of hyperspectral images, and a sparse constraint is imposed on each layer to make the unmixing results more consistent with the physical mechanism of mixed spectral pixels. With these constraints, the proposed method established a spectral–spatial joint tensor decomposition model that represents not only the local neighborhood similarity but also the heterogeneity of adjacent edges. Experiments on simulated data and real hyperspectral data demonstrate the effectiveness of the proposed method.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Hyperspectral Image Classification With Contrastive Graph Convolutional
           Network

    • Free pre-print version: Loading...

      Authors: Wentao Yu;Sheng Wan;Guangyu Li;Jian Yang;Chen Gong;
      Pages: 1 - 15
      Abstract: Recently, graph convolutional network (GCN) has been widely used in hyperspectral image (HSI) classification due to its satisfactory performance. However, the number of labeled pixels is very limited in HSI, and thus, the available supervision information is usually insufficient, which will inevitably degrade the representation ability of most existing GCN-based methods. To enhance the feature representation ability, in this article, a GCN model with contrastive learning is proposed to explore the supervision signals contained in both spectral information and spatial relations, which is termed contrastive GCN (ConGCN), for HSI classification. First, in order to mine sufficient supervision signals from spectral information, a semisupervised contrastive loss function is utilized to maximize the agreement between different views of the same node or the nodes from the same land cover category. Second, to extract the precious yet implicit spatial relations in HSI, a graph generative loss function is leveraged to explore supplementary supervision signals contained in the graph topology. In addition, an adaptive graph augmentation technique is designed to flexibly incorporate the spectral–spatial priors of HSI, which helps facilitate the subsequent contrastive representation learning. The extensive experimental results on six typical benchmark datasets firmly demonstrate the effectiveness of the proposed ConGCN in both qualitative and quantitative aspects.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Hybrid Fully Connected Tensorized Compression Network for Hyperspectral
           Image Classification

    • Free pre-print version: Loading...

      Authors: Heng-Chao Li;Zhi-Xin Lin;Tian-Yu Ma;Xi-Le Zhao;Antonio Plaza;William J. Emery;
      Pages: 1 - 16
      Abstract: Deep learning models, such as convolutional neural networks (CNNs), have made significant progress in hyperspectral image (HSI) classification. However, these models require a large number of parameters, which occupy a lot of storage space and suffer from overfitting, thus resulting in performance loss. To solve the above problems, in this article, we propose a new compression network [namely, a Hybrid Fully Connected Tensorized Compression Network (HybridFCTCN)] by considering the high dimensionality of HSI data. First, using the low-rank fully connected tensor network decomposition (FCTND), three novel units, i.e., FCTN-FC, FCTNConv2D, and FCTNConv3D, are designed to compress the weight tensor of standard fully connected (FC) layer and kernel tensor of convolutional layer, reducing their parameters. In the novel units, the intrinsic correlation of the decomposed factors is adequately exploited by the FC structures, which enhances their feature extraction and classification abilities. Then, benefiting from the hybrid network backbone composed of the FCTNConv3D and FCTNConv2D units, HybridFCTCN can extract more discriminative features with fewer parameters, while it has great generalization capability and robustness, enabling better HSI classification. Finally, the rank of above-designed units is defined, and its determination is discussed to facilitate the application of the proposed model. Extensive experiments on three widely used HSI datasets reveal that the proposed model achieves state-of-the-art classification performance for different training sample sizes with a very small number of parameters.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Bathymetry Retrieval Algorithm Based on Hyperspectral Features of Pure
           Water Absorption From 570 to 600 nm

    • Free pre-print version: Loading...

      Authors: Zhongqiang Wu;Bangyi Tao;Zhihua Mao;Haiqing Huang;
      Pages: 1 - 19
      Abstract: Current efforts for improving the hyperspectral optimization processing exemplar (HOPE) model include further testing of remote sensing reflectance ( $R_{mathrm {rs}}$ ) features containing useful information for bathymetry retrieval via the minimization of the interference stemming from the variability in inherent optical properties and benthic reflectance. In this article, we detected a novel feature originating from the pure water absorption within the narrow spectral region of 570–600 nm. In most coastal regions of clear water in coral reefs, for example, in a coral reefs environment, pure water accounts for the majority of the total absorption in this spectral range. In addition to the depth variation, the spectral behavior of $R_{mathrm {rs}}$ (570–600) is primarily dominated by a steep increase in pure water absorption with wavelength, whereas the influence of other optical properties, such as phytoplankton/colored dissolved organic matter (CDOM) absorption, particle backscattering, and benthic reflectance, can be simplified using the spectrally constant shape model. An HOPE pure water (HOPE-PW) algorithm using this feature was developed based on $R_{mathrm {rs}}$ measurements with a spectral resolution of near 3.5 nm, in which only four uncertainties must be resolved. The validation from light detection and ranging (LiDAR) data and comparison with HOPE-bottom reflectance unmixing computation of the environment (BRUCE) using portable remote imaging radiometer (PRISM) data at 15 sites located in five distinct regions of Palau, Guam, Great Barrier Reef, Hawaiian Islands, and Florida Key confirmed that the HOPW-PW algorithm yielded a considerable performance and provided adequate transferability to other sit-s with varying bottom and water environments. Furthermore, the sensitivity analysis based on Hydrolight-simulated datasets was carried out and showed that HOPE-PW was less affected by variation of bottom types, but still had some limitations in retrieving water optical properties.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Hyperspectral Band Selection via Difference Between Intergroups

    • Free pre-print version: Loading...

      Authors: Shuying Li;Baidong Peng;Long Fang;Qiang Zhang;Lei Cheng;Qiang Li;
      Pages: 1 - 10
      Abstract: Various methods are proposed to reduce the dimensions of hyperspectral image (HSI) by band selection in recent years. Most methods select one band from each group to construct a band subset. However, the redundancy in the selected bands from different groups is neglected. Furthermore, the researchers do not pay enough attention to how many bands are appropriated for selection. To solve these issues, we propose a hyperspectral band selection method via difference between intergroups (DIG), which includes grouping strategy and ranking strategy. Specifically, the grouping strategy adopts intragroup similarity to reasonably distribute all partitioning point positions. The similarity of bands within the same group is significantly improved. For the ranking strategy, it not only takes into account the knowledge and intragroup similarity of bands, but also evaluates the differences between each band and other intergroup bands. The redundancy in band subset is reduced sufficiently. To accurately obtain the optimal number of bands, an evaluation function is designed to measure the information content and redundancy in various band subsets. Experimental results from different aspects show that the proposed model has a large performance advantage on three public datasets.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Positive Feedback Spatial-Spectral Correlation Network Based on Spectral
           Slice for Hyperspectral Image Classification

    • Free pre-print version: Loading...

      Authors: Cuiping Shi;Haiyang Wu;Liguo Wang;
      Pages: 1 - 17
      Abstract: The emergence of convolutional neural networks (CNNs) has greatly promoted the development of hyperspectral image classification (HSIC). However, some serious problems are the lack of label samples in hyperspectral images (HSIs), and the spectral characteristics of different objects in HSIs are sometimes similar among classes. These problems hinder the improvement of HSIC performance. To this end, in this article, a positive feedback spatial-spectral correlation network based on spectral interclass slicing (PFSSC_SICS) is proposed. First, a spectral interclass slicing (SICS) strategy is designed, which can remove similar spectral signature between classes and reduce the impact of similar spectral signature of different classes on HSIC performance. Second, in order to solve the impact of the lack of labeled samples on HSIC, a positive feedback (PF) mechanism and a spatial-spectral correlation (SSC) module are introduced to extract deeper and more features. Finally, the experimental results show that the classification performance of the PFSSC_SICS is far exceed than that of some state-of-the-art methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Large-Scale Hyperspectral Image Restoration via a Superpixel Distributed
           Algorithm Based on Graph Signal Processing

    • Free pre-print version: Loading...

      Authors: Wanyuan Cai;Junzheng Jiang;Jiang Qian;
      Pages: 1 - 17
      Abstract: Hyperspectral image (HSI) is often disturbed by various kinds of noise, which brings great challenges to subsequent applications. Many of the existing restoration algorithms do not scale well for HSI with large size. This article proposes a novel mixed-noise removal method for HSI with large size, by leveraging the superpixel segmentation-based technology and distributed algorithm based on graph signal processing (GSP). First, the underlying structure of the HSI is modeled by a two-layer architecture graph. The upper layer, called skeleton graph, is a rough graph constructed using the modified $k$ -nearest-neighborhood algorithm and its nodes correspond to a series of superpixels formed by HSI segmentation. The skeleton graph can efficiently characterize the intercorrelations between superpixels, while preserving the boundary information and reducing the computational complexity. The lower layer, called detailed graph consisting of a series of local graphs which are constructed to model the similarities between pixels. Second, based on the two-layer graph architecture, the HSI restoration problem is formulated as a series of optimization problems each of which resides on a subgraph. In each optimization problem, a graph Laplacian regularization (GLR) is defined and incorporated into a low-rank (LR)-based model. Third, a novel distributed algorithm is tailored for the restoration problem, using the information interaction between the nodes of skeleton graph and subgraphs. Numerical experiments conducted on both synthetic and real-world datasets demonstrate the effectiveness of the proposed restoration algorithm compared with existing methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Spectral–Spatial Morphological Attention Transformer for
           Hyperspectral Image Classification

    • Free pre-print version: Loading...

      Authors: Swalpa Kumar Roy;Ankur Deria;Chiranjibi Shah;Juan M. Haut;Qian Du;Antonio Plaza;
      Pages: 1 - 15
      Abstract: In recent years, convolutional neural networks (CNNs) have drawn significant attention for the classification of hyperspectral images (HSIs). Due to their self-attention mechanism, the vision transformer (ViT) provides promising classification performance compared to CNNs. Many researchers have incorporated ViT for HSI classification purposes. However, its performance can be further improved because the current version does not use spatial–spectral features. In this article, we present a new morphological transformer (morphFormer) that implements a learnable spectral and spatial morphological network, where spectral and spatial morphological convolution operations are used (in conjunction with the attention mechanism) to improve the interaction between the structural and shape information of the HSI token and the CLS token. Experiments conducted on widely used HSIs demonstrate the superiority of the proposed morphFormer over the classical CNN models and state-of-the-art transformer models. The source will be made available publicly at https://github.com/mhaut/morphFormer.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • PSRT: Pyramid Shuffle-and-Reshuffle Transformer for Multispectral and
           Hyperspectral Image Fusion

    • Free pre-print version: Loading...

      Authors: Shang-Qi Deng;Liang-Jian Deng;Xiao Wu;Ran Ran;Danfeng Hong;Gemine Vivone;
      Pages: 1 - 15
      Abstract: A Transformer has received a lot of attention in computer vision. Because of global self-attention, the computational complexity of Transformer is quadratic with the number of tokens, leading to limitations for practical applications. Hence, the computational complexity issue can be efficiently resolved by computing the self-attention in groups of smaller fixed-size windows. In this article, we propose a novel pyramid Shuffle-and-Reshuffle Transformer (PSRT) for the task of multispectral and hyperspectral image fusion (MHIF). Considering the strong correlation among different patches in remote sensing images and complementary information among patches with high similarity, we design Shuffle-and-Reshuffle (SaR) modules to consider the information interaction among global patches in an efficient manner. Besides, using pyramid structures based on window self-attention, the detail extraction is supported. Extensive experiments on four widely used benchmark datasets demonstrate the superiority of the proposed PSRT with a few parameters compared with several state-of-the-art approaches. The related code is available at https://github.com/Deng-shangqi/PSRT.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Deep Hierarchical Pyramid Network With High- Frequency -Aware Differential
           Architecture for Super-Resolution Mapping

    • Free pre-print version: Loading...

      Authors: Da He;Yanfei Zhong;
      Pages: 1 - 15
      Abstract: Super-resolution mapping (SRM) is a way to solve the mixed-pixel problem in urban land use/land cover caused by the limited spatial-resolving ability of satellite sensors, through resolution enhancement of the classification map. Recently, deep learning-based super-resolution mapping (DLSM) networks have been boomed, which can automatically learn a mapping pattern from low-resolution (LR) image to high-resolution (HR) land cover distribution to alleviate mixed-pixel problem. However, the urban compositions like buildings, trees, and roads exhibit a multiscale distribution with different size or orientation, which makes the traditional single-scale DLSM failed for an appropriate recognition. In addition, the urban compositions also show significant spatial heterogeneity with irregular distribution and intricate morphological shape, which are difficult to learn by simple convolutional layer. Therefore, it is necessary to explore the cue of these distribution characteristic to constrain the learning behavior of the network for better detail restoration. In this article, a deep hierarchical pyramid sub-pixel mapping network (HiSMNet) with high-frequency-aware differential architecture is proposed, which establishes an HP architecture to achieve explicit multiscale supervision of the feature map and prompt the network to learn a multiscale representation. In addition, a differential architecture is designed to enforce the network to intensify the learning of the high-frequency details. The validation experiments demonstrate that HiSMNet achieves superior performances in detailed delineation and outperformed the state-of-the-art DLSM models by up to 10% in terms of overall accuracy.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Deep Hyperspectral and Multispectral Image Fusion via Probabilistic Matrix
           Factorization

    • Free pre-print version: Loading...

      Authors: Baihong Lin;Yulan Guo;
      Pages: 1 - 14
      Abstract: Deep learning methods are popular for hyperspectral and multispectral image (HSI-MSI) fusion to obtain a high-resolution HSI. However, most of them are unsatisfactory due to limited generalization ability and poor interpretability. This article proposes a highly interpretable deep HSI-MSI fusion method based on probabilistic matrix factorization (PMF) under the Bayesian framework. In the proposed method, an HSI is factorized into two matrices, namely, the Gaussian-prior-regularized spectral matrix and the deep-prior-regularized abundance matrix. Then, we split the optimization process into two meaningful iterative updating steps: updating the spectral matrix based on least-squares estimation and updating the abundance matrix based on a convolutional neural network (CNN)-based Gaussian denoiser for 2-D gray images. To improve the generalization ability, we provide solutions for selections of hyperparameters, CNN-based denoiser architecture, and training strategy. Using the given solutions, the proposed fusion method can be trained with 2-D images once and then used to fuse different types of HSI and MSI excellently. Experiments on three datasets demonstrate that the proposed fusion method has good fusion performance and high generalization ability compared with other state-of-the-art methods. The source code will be available at https://github.com/KevinBHLin/.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Pyramidal Multiscale Convolutional Network With Polarized Self-Attention
           for Pixel-Wise Hyperspectral Image Classification

    • Free pre-print version: Loading...

      Authors: Haimiao Ge;Liguo Wang;Moqi Liu;Xiaoyu Zhao;Yuexia Zhu;Haizhu Pan;Yanzhong Liu;
      Pages: 1 - 18
      Abstract: In recent years, pixel-wise hyperspectral image (HSI) classification has received growing attention in the field of remote sensing. Plenty of spectral–spatial convolutional neural network (CNN) methods with diverse attention mechanisms have been proposed for HSI classification due to the attention mechanisms being able to provide more flexibility over standard convolutional blocks. However, it remains a challenge to effectively extract multiscale features of high-resolution HSI in a real-world complex environment. In this article, we propose a pyramidal multiscale spectral–spatial convolutional network with polarized self-attention for pixel-wise HSI classification. It contains three stages: channel-wise feature extraction network, spatial-wise feature extraction network, and classification network, which are used to extract spectral features, extract spatial features, and generate classification results, respectively. Pyramidal convolutional blocks and polarized attention blocks are combined to extract spectral and spatial features of HSI. Furthermore, residual aggregation and one-shot aggregation are employed to better converge the network. The experimental results on several public HSI datasets demonstrate that the proposed network outperforms other related methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • A Global-to-Local Evolutionary Algorithm for Hyperspectral Endmember
           Extraction

    • Free pre-print version: Loading...

      Authors: Fan Cheng;Naikun Chen;Chao Wang;Qijun Wang;Bo Du;
      Pages: 1 - 17
      Abstract: Recently, evolutionary algorithms (EAs) have shown their promising performance in solving the hyperspectral endmember extraction (EE) task. Despite that, most of the existing EA-based EE algorithms mainly take advantage of the global search capability of evolutionary computation. A few of them focus on the hyperspectral EE task itself, which is a sparse large-scale problem with constraint. To fill the gap, in this article, a global-to-local EA (GL-EA) is proposed, where the global and local search is performed sequentially to extract the endmembers effectively. Specifically, in the first global search stage, two complementary solution generation strategies, including asymmetric flip-based solution generation and spectral angle distance (SAD)-based solution repair, are designed, with which the sparse large-scale search space of hyperspectral EE is fully explored and the endmembers that satisfy the constraint could be achieved. Then, in the second stage, a perturbation-based local search is suggested, which further enhances the quality of the obtained endmembers. In addition, an endmember repetition-based solution selection strategy is also developed for both global and local search stages, by using which good solutions can be selected effectively during the evolution. Experimental results on different hyperspectral datasets demonstrate that when compared with the state-of-the-art EE algorithms, the proposed GL-EA could extract the endmembers with higher quality.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • BS3LNet: A New Blind-Spot Self-Supervised Learning Network for
           Hyperspectral Anomaly Detection

    • Free pre-print version: Loading...

      Authors: Lianru Gao;Degang Wang;Lina Zhuang;Xu Sun;Min Huang;Antonio Plaza;
      Pages: 1 - 18
      Abstract: Recent years have witnessed the flourishing of deep learning-based methods in hyperspectral anomaly detection (HAD). However, the lack of available supervision information persists throughout. In addition, existing unsupervised learning/semisupervised learning methods to detect anomalies utilizing reconstruction errors not only generate backgrounds but also reconstruct anomalies to some extent, complicating the identification of anomalies in the original hyperspectral image (HSI). In order to train a network able to reconstruct only background pixels (instead of anomalous pixels), in this article, we propose a new blind-spot self-supervised learning network (called BS3LNet) that generates training patch pairs with blind spots from a single HSI and trains the network in self-supervised fashion. The BS3LNet tends to generate high reconstruction errors for anomalous pixels and low reconstruction errors for background pixels due to the fact that it adopts a blind-spot architecture, i.e., the receptive field of each pixel excludes the pixel itself and the network reconstructs each pixel using its neighbors. The above characterization suits the HAD task well, considering the fact that spectral signatures of anomalous targets are significantly different from those of neighboring pixels. Our network can be considered a superb background generator, which effectively enhances the semantic feature representation of the background distribution and weakens the feature expression for anomalies. Meanwhile, the differences between the original HSI and the background reconstructed by our network are used to measure the degree of the anomaly of each pixel so that anomalous pixels can be effectively separated from the background. Extensive experiments on two synthetic and three real datasets reveal that our BS3LNet is competitive with regard to other state-of-the-art approaches.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Iterative Spectral–Spatial Hyperspectral Anomaly Detection

    • Free pre-print version: Loading...

      Authors: Chein-I Chang;Chien-Yu Lin;Pau-Choo Chung;Peter Fuming Hu;
      Pages: 1 - 30
      Abstract: Anomaly detection (AD) requires spectral and spatial information to differentiate anomalies from their surrounding data samples. To capture spatial information, a general approach is to utilize local windows in various forms to adapt local characteristics of the background (BKG) from which unknown anomalies can be detected. This article develops a new approach, called iterative spectral–spatial hyperspectral AD (ISSHAD), which can improve an anomaly detector in its performance via an iterative process. Its key idea is to include an iterative process that captures spectral and spatial information from AD maps (ADMaps) obtained in previous iterations and feeds these anomaly maps back to the current data cube to create a new data cube for the next iteration. To terminate the iterative process, a Tanimoto index (TI)-based automatic stopping rule is particularly designed. Three types of spectral and spatial information, ADMaps, foreground map (FGMap), and spatial filtered map (SFMap), are introduced to develop seven various versions of ISSHAD. To demonstrate its full utilization in improving AD performance, a large number of extensive experiments are performed for ISSHAD along with its detailed comprehensive analysis among several most recently developed anomaly detectors, including classic, dual-window-based, low-rank representation model-based, and tensor-based AD methods for validation.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Matrix Factorization With Framelet and Saliency Priors for Hyperspectral
           Anomaly Detection

    • Free pre-print version: Loading...

      Authors: Xiangfei Shen;Haijun Liu;Jing Nie;Xichuan Zhou;
      Pages: 1 - 13
      Abstract: Hyperspectral anomaly detection aims to separate sparse anomalies from low-rank background components. A variety of detectors have been proposed to identify anomalies, but most of them tend to emphasize characterizing backgrounds with multiple types of prior knowledge and limited information on anomaly components. To tackle these issues, this article simultaneously focuses on two components and proposes a matrix factorization method with framelet and saliency priors to handle the anomaly detection problem. We first employ a framelet to characterize nonnegative background representation coefficients, as they can jointly maintain sparsity and piecewise smoothness after framelet decomposition. We then exploit saliency prior knowledge to measure each pixel’s potential to be an anomaly. Finally, we incorporate the pure pixel index (PPI) with Reed-Xiaoli’s (RX) method to possess representative dictionary atoms. We solve the optimization problem using a block successive upper-bound minimization (BSUM) framework with guaranteed convergence. Experiments conducted on benchmark hyperspectral datasets demonstrate that the proposed method outperforms some state-of-the-art anomaly detection methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Adaptive Reference-Related Graph Embedding for Hyperspectral Anomaly
           Detection

    • Free pre-print version: Loading...

      Authors: Yubo Ma;Siyu Cai;Jie Zhou;
      Pages: 1 - 14
      Abstract: Graph embedding (GE) provides an effective way to reveal the intrinsic feature of high-dimensional data on the foundation of preserving topological properties. Under the framework of GE, the hyperspectral image can be represented by a weighted graph, where pixels and similarities among them are treated as vertices and edge weights, respectively. In this article, an adaptive reference-related GE (ARGE) method is proposed to efficaciously obtain the low-dimensional feature and improve computational efficiency. The ARGE method is composed of two primary processes. The key to connecting these two processes is the reference vertices set, which is the abstraction of graph topological features. First, the reference vertices are adaptively selected through a three-step adaptive reference set selection (ARSS) algorithm. Second, the original high-dimensional graph is embedded as a low-dimensional graph through preserving the reference-related structure. Specifically, the pairwise similarities between vertices and reference vertices are preserved in embedding space. In addition, a new hybrid dissimilarity measure of Rao distance and spectral information divergence (RD-SID) is designed to depict the spectral difference between pixels. To evaluate the effectiveness of the proposed method, the obtained low-dimensional feature is fed into the anomaly detector to detect anomalous pixels. The experimental results on five real and one synthetic hyperspectral datasets demonstrate the superiority of the proposed ARGE method over the compared feature extraction methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • Diffused Convolutional Neural Network for Hyperspectral Image
           Super-Resolution

    • Free pre-print version: Loading...

      Authors: Sen Jia;Shuangzhao Zhu;Zhihao Wang;Meng Xu;Weixi Wang;Yujuan Guo;
      Pages: 1 - 15
      Abstract: With the rapid development of deep convolutional neural networks (CNNs), super-resolution (SR) in hyperspectral image (HSI) has achieved good results. Current methods generally use 2-D convolution for feature extraction, but they cannot effectively extract spectral information. Although 3-D convolution can better characterize feature structure of HSI, it will lead to parameter redundancy, model complexity, and severe memory shortage. To address the above problems, we propose a new HSI SR method, named diffused CNN (DCNN). Specifically, spectral convolutions have been added into the enhanced convolutional neural (ECN) block, and a series of spectral convolutions are introduced in the residual network to learn features in the channel direction of different depths. Furthermore, histogram of oriented gradient (HOG) and local binary pattern (LBP) are used to retain the shape and texture information of the image, respectively, which can well represent the spatial structure of the object. To effectively make use of the extracted shallow and deep features, a feature fusion strategy is used to reinforce the reconstruction efficiency. Besides, an image enhancement module has been developed to diffuse the SR image into the image space. Extensive evaluations and comparisons show that our DCNN approach can not only recover the HSI data with richer details but also achieve superiority over several state-of-the-art methods.
      PubDate: 2023
      Issue No: Vol. 61 (2023)
       
  • First-Order Smoothing-Based Deep Graph Network for Hyperspectral Image
           Classification

    • Free pre-print version: Loading...

      Authors: Yizhen Li;Yanwen Chong;Shaoming Pan;Yun Ding;
      Pages: 1 - 16
      Abstract: Although graph convolutional network (GCN) has achieved remarkable success in hyperspectral image (HSI) classification, most existing GCN-based approaches have failed to realize a deep network structure due to the oversmoothing problem. This problem largely limits the expression ability and feature extraction ability of GCN and hampers GCN’s capacity to model long-range relationships between samples in hyperspectral (HS) scenes. Moreover, there is a lack of theoretical analysis in those works that constructed deep GCN for HSI classification to illustrate how they overcome the oversmoothing problem. Aside from this, the characteristics and complexity of HSI are often neglected when constructing deep GCN models in HSI classification. To address these problems, a novel deep graph network based on first-order smoothing is proposed for HSI classification. Specifically, a local and global topologically consistent graph is constructed to thoroughly explore the union between fine pixel information and semantic superpixel information. Subsequently, a novel propagation procedure is proposed to address the oversmoothing problem. We creatively build a residual connection to the first layer to emphasize the feature information aggregated from the first-order neighborhood, which adds node features that have not yet become indistinguishable into deep layer, and at the same time, it can be considered as a correction to the original pixels affected by spectral variation in the input graph. Finally, we demonstrate how first-order smoothing-based deep graph network (FSDGN) can slow down the conver