 Subjects -> INSTRUMENTS (Total: 63 journals)
|
|
|
- Generating Virtual Training Labels for Crop Classification from Fused
Sentinel-1 and Sentinel-2 Time Series-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Convolutional neural networks (CNNs) have shown results superior to most traditional image understanding approaches in many fields, incl. crop classification from satellite time series images. However, CNNs require a large number of training samples to properly train the network. The process of collecting and labeling such samples using traditional methods can be both, time-consuming and costly. To address this issue and improve classification accuracy, generating virtual training labels (VTL) from existing ones is a promising solution. To this end, this study proposes a novel method for generating VTL based on sub-dividing the training samples of each crop using self-organizing maps (SOM), and then assigning labels to a set of unlabeled pixels based on the distance to these sub-classes. We apply the new method to crop classification from Sentinel images. A three-dimensional (3D) CNN is utilized for extracting features from the fusion of optical and radar time series. The results of the evaluation show that the proposed method is effective in generating VTL, as demonstrated by the achieved overall accuracy (OA) of 95.3% and kappa coefficient (KC) of 94.5%, compared to 91.3% and 89.9% for a solution without VTL. The results suggest that the proposed method has the potential to enhance the classification accuracy of crops using VTL.
PubDate: 2023-09-26
- Assessing the Physical and Chemical Characteristics of Marine Mucilage
Utilizing In-Situ and Remote Sensing Data (Sentinel-1, -2, -3)-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract In spring 2021, mucilage, also known as “sea snot” or “sea saliva” has been intensely observed in the Sea of Marmara and has reached a threatening level. Due to the declining water quality, many marine organisms have perished, the fishing industry and tourism have been adversely affected. In this paper, a detailed investigation was carried out to assess the effects of mucilage phenomenon on the seawater quality, sea surface temperature, and backscattered radar signal power in two separate mucilage-covered areas in the Sea of Marmara. The quality of the mucilage-covered seawater was investigated by calculating physico-chemical parameters such as sea surface temperature, electrical conductivity, the potential of hydrogen, suspended solids, dissolved oxygen concentration, and chlorophyll-a in the water samples taken. With in-situ measurements, the spectral responses of intense and middle-intense mucilage were determined by a full-range spectroradiometer and compared with the spectral signature of clean seawater. Furthermore, utilizing space-borne synthetic aperture radar (SAR) and optical images of Sentinel-1, Sentinel-2 and Sentinel-3, the effects of mucilage on spectral reflectance, radar signal backscattering, and sea surface temperature were investigated depending on its intensity. The results of in-situ measurements and laboratory analyses showed considerable effects of mucilage on water quality. The space-borne analyses demonstrated that middle-intense and intense mucilage cause approximately 0.5 and 1-decibel decrease in backscattered radar signal power against clean water. In terms of sea surface temperature, the differences between clean seawater and middle-intense and intense mucilage areas were estimated as 1.05–2.25 °C, respectively.
PubDate: 2023-09-19
- Comparative Analysis of Multispectral and Hyperspectral Imagery for
Mapping Sugarcane Varieties-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Mapping different varieties of sugarcane is vital for estimating yields and assessing crop damage risks. Different biophysical and chemical characteristics of sugarcane varieties at different ages lead to distinct electromagnetic spectral behaviors. Remote sensing images can be useful in mapping sugarcane varieties. Multispectral imagery was evaluated and compared with hyperspectral imagery for mapping different varieties of sugarcane crops in this study. A set of field data on various varieties of sugarcane farms in Khuzestan Province, Iran was used alongside satellite imagery from Landsat ETM+ and EO-1 Hyperion sensors to achieve this objective. In order to determine the most optimal spectral bands from a Hyperion hyperspectral image for classification of sugarcane varieties, particle swarm optimization (PSO) was employed as the primary optimization algorithm. Several spectral indices and a principal component analysis (PCA) were also applied for extracting surface biophysical properties. The training dataset was then used to classify hyperspectral and multispectral images at a variety and variety-age scales using various supervised classification methods. Further comparisons with field data were conducted to determine the accuracy of classification results per classification method, as well as for both types of images. According to the study findings, selecting ETM+ and optimal Hyperion bands to map sugarcane varieties was most effective with 69 and 81% accuracy, respectively. As a result, accuracy was further improved to 76 and 82%, respectively. This was due to the incorporation of vegetation indices, PCA components, and soil salinity indices contributing to these improvements. In addition, using sugarcane age as a classification feature resulted in an increase in accuracy of 3 and 6% for ETM+ and Hyperion images, respectively. In the study, the support vector machine (SVM) demonstrated the highest accuracy among different classification techniques. These results indicate the value of employing surface biophysical properties in conjunction with multispectral images. As an alternative to hyperspectral images for agricultural varieties classification, this approach can be suggested.
PubDate: 2023-09-06
- Guiding Deep Learning with Expert Knowledge for Dense Stereo Matching
-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Dense depth information can be reconstructed from stereo images using conventional hand-crafted as well as deep learning-based approaches. While deep-learning methods often show superior results compared to hand-crafted ones, they commonly learn geometric principles underlying the matching task from scratch and neglect that these principles have already been intensively studied and were considered explicitly in various models with great success in the past. In consequence, a broad range of principles and associated features need to be learned, limiting the possibility to focus on important details to also succeed in challenging image regions, such as close to depth discontinuities, thin objects and in weakly textured areas. To overcome this limitation, in this work, a hybrid technique, i.e., a combination of conventional hand-crafted and deep learning-based methods, is presented, addressing the task of dense stereo matching. More precisely, the input RGB stereo images are supplemented by a fourth image channel containing feature information obtained with a method based on expert knowledge. In addition, the assumption that edges in an image and discontinuities in the corresponding depth map coincide is modeled explicitly, allowing to predict the probability of being located next to a depth discontinuity per pixel. This information is used to guide the matching process and helps to sharpen correct depth discontinuities and to avoid the false prediction of such discontinuities, especially in weakly textured areas. The performance of the proposed method is investigated on three different data sets, including studies on the influence of the two methodological components as well as on the generalization capability. The results demonstrate that the presented hybrid approach can help to mitigate common limitations of deep learning-based methods and improves the quality of the estimated depth maps.
PubDate: 2023-07-28
- Crowd-aware Thresholded Loss for Object Detection in Wide Area Motion
Imagery-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Detecting objects in Wide Area Motion Imagery (WAMI), an essential task for many practical applications, is particularly challenging in crowded scenes, such as areas with heavy traffic, since pixel resolutions of objects and ground sampling distance are highly compromised, and different factors disrupt visual signals. To address this challenge, we design a framework that combines preprocessing operations and deep detectors. To train deep networks for detection in WAMI for improved performance in especially crowded areas, we propose a novel crowd-aware thresholded loss (CATLoss) function. Moreover, we introduce a hard sampling mining method to strengthen the discriminative ability of the proposed solution. Additionally, we extend prior networks used in the literature using novel spatio-temporal cascaded architectures to incorporate more contextual information without introducing additional parameters. Overall, our approach is causal, more generalizable, and more robust even in reduced spatial sizes. On the WPAFB-2009 dataset, we show that our solution performs better than or on par with state-of-the-art without introducing any computational complexity during inference. The code and trained models will be released at (https://github.com/poyrazhatipoglu/CATLoss).
PubDate: 2023-07-24
- Spatial Downscaling of Snow Water Equivalent Using Machine Learning
Methods Over the Zayandehroud River Basin, Iran-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Snow cover is an informative indicator of climate change and surface hydrological cycles. Despite its essential accurate dynamic measurement (i.e., accumulation, erosion, and runoff), it is poorly known, particularly in mountainous regions. Since passive microwave sensors can contribute to obtaining information about snowpack volume, microwave brightness temperatures (BT) have long been used to assess spatiotemporal variations in snow water equivalent (SWE). However, SWE is greatly influenced by geographic location, terrain parameters/covers, and BT differences, and thus, the low spatial resolution of existing SWE products (i.e., the coarse resolution of AMSR-based products) leads to less satisfactory results, especially in regions with complex terrain conditions, strong seasonal transitions and, great spatiotemporal heterogeneity. A novel multifactor SWE downscaling algorithm based on the support vector regression (SVR) technique has been developed in this study for the Zayandehroud River basin. Thereby, passive microwave BT, location (latitude and longitude), terrain parameters (i.e., elevation, slope, and aspect), and vegetation cover serve as model input data. Evaluation of downscaled SWE estimates against ground-based observations demonstrated that when moving into higher spatial resolution, not only was there no significant decrease in accuracy, but a 4% increase was observed. In addition, this study suggests that integrating passive microwave remote sensing data with other auxiliary data can lead to a more efficient and effective algorithm for retrieving SWE with appropriate spatial resolution over various scales.
PubDate: 2023-07-21
- Evaluation of InSAR Tropospheric Correction Methods over North-West Iran
-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Synthetic Aperture Radar (SAR) interferometric measurements are highly affected by tropospheric delay caused by different troposphere conditions between SAR acquisitions. Tropospheric correction is even more critical in measuring the small surface deformation such as interseismic slips over tectonic faults. Various approaches from which some of them are based on auxiliary data and weather models and the other ones are relied on only phase measurements have been already proposed for tropospheric correction. The capability of these methods, however, has not been investigated over areas such as Iran which are suffering from insufficient data and inaccurate performance of weather models. Hence, as the main goal in this study, a comprehensive statistical comparison between tropospheric correction methods is performed over the North-West Iran which is subject to high tectonic activity. The correction methods based on spectrometer data (MERIS), weather models (MERRA-2 and ERA-Interim), topography-dependent functions (linear and power-law relations) and a combination of low- and high-pass (L/H) filters are applied on single-master interferometric phase. The persistent scatterer time series analysis using 12 Envisat ASAR raw data in two different frames spanning between 2004 and 2008 is applied to estimate the interseismic deformation. The results show that the L/H filters which are based on only phase measurements is more able to reduce the phase spatial variations. None of the other tropospheric correction methods consistently reduced tropospheric signals over different times due to lack of accurate weather model and auxiliary data in Iran. The superiority of the combined filters over other approaches is that no auxiliary data is required. More importantly, the deformation due to the interseismic slip over the active faults in the study area is considerably preserved after the tropospheric correction using L/H filters while this is not the case for other correction approaches.
PubDate: 2023-07-19
DOI: 10.1007/s41064-023-00250-2
- Reports
-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
PubDate: 2023-07-17
DOI: 10.1007/s41064-023-00251-1
- Reports
-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
PubDate: 2023-06-21
DOI: 10.1007/s41064-023-00245-z
- A Metaheuristic Optimization-Based Solution to MTF-GLP-Based Pansharpening
-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract In recent years, the pansharpening strategies employing the Generalized Laplacian Pyramid (GLP) based on Gaussian filters that match the Modulation Transfer Function (MTF) of the source multispectral (MS) sensor have attracted attention in remote sensing community. The MTF-GLP-based pansharpening methods differ from each other in the way they obtain the injection coefficients, which are used to transfer the spatial details of the source panchromatic (PAN) image into the source MS image. Investigation of the pansharpening literature showed that the MTF-GLP-based pansharpening strategies generally estimate the injection coefficients using statistics-based deterministic approaches, which leads to a difficulty in identifying the non-linear relationship between the source MS and PAN data. Hence, this study proposes a metaheuristic optimization-based solution to this problem. The proposed method estimates the optimum injection coefficients through the Multi-Objective Symbiotic Organism Search (MOSOS) algorithm, which has been proven to efficiently find the optimum solutions in very complex search spaces. The success of the presented method was qualitatively and quantitatively tested on four test sites against several widely used pansharpening techniques. The experiments revealed that the presented approach did not only outperform some of the commonly used MTF-GLP-based methods, but also some of the other Multiresolution Analysis (MRA)-based, component substitution (CS)-based, deep learning (DL)-based, and variational optimization (VO)-based pansharpening methods.
PubDate: 2023-06-16
DOI: 10.1007/s41064-023-00248-w
- Impact of Drone Regulations on Drone Use in Geospatial Applications and
Research: Focus on Visual Range Conditions, Geofencing and Privacy
Considerations-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract The European Aviation Safety Agency (EASA) laid down the new EU drone regulations in December 2020, which are seen as a development step towards enabling a reliable legal framework for drone operation and use. At the same time, it is still not clear how the new rules will influence drone usage in various drone-based applications and sectors. Therefore, the paper aims to discuss and analyse how recent rules may influence drone use and affect its economic viability with the focus on the geospatial sector. The discussion of rules’ impact is achieved based on three key items: visual range limitations, geofencing systems (virtual geographic boundary around specific areas of interest) and, finally, the effect of privacy considerations. To enrich the discussion and get more insight into rules’ impact, important issues from different technical and economic aspects were distributed in a questionnaire to collect data from participants who are actively developing or using the drone technology. The questionnaire results revealed that the majority of participants (75%) were of the opinion that it is somehow difficult to judge how the new regulations will help/hinder the drone use at least in the current stage of rules implementation. In addition, there is a tendency in conducting drone flights under visual line of sight conditions even if there are difficulties in keeping a continuous contact with the drone during flight operations. Finally, the results pointed out that 44% of participants faced problems with privacy issues that affected their flight missions, and as a consequence some of the projects even got cancelled.
PubDate: 2023-06-15
DOI: 10.1007/s41064-023-00246-y
- Comparison of an Optimised Multiresolution Segmentation Approach with Deep
Neural Networks for Delineating Agricultural Fields from Sentinel-2 Images
-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Effective monitoring of agricultural lands requires accurate spatial information about the locations and boundaries of agricultural fields. Through satellite imagery, such information can be mapped on a large scale at a high temporal frequency. Various methods exist in the literature for segmenting agricultural fields from satellite images. Edge-based, region-based, or hybrid segmentation methods are traditional methods that have widely been used for segmenting agricultural fields. Lately, the use of deep neural networks (DNNs) for various tasks in remote sensing has been gaining traction. Therefore, to identify the optimal method for segmenting agricultural fields from satellite images, we evaluated three state-of-the-art DNNs, namely Mask R-CNN, U-Net, and FracTAL ResUNet against the multiresolution segmentation (MRS) algorithm, which is a region-based and a more traditional segmentation method. Given that the DNNs are supervised methods, we used an optimised version of the MRS algorithm based on supervised Bayesian optimisation. Monotemporal Sentinel-2 (S2) images acquired in Lower Saxony, Germany were used in this study. Based on the agricultural parcels declared by farmers within the European Common Agricultural Policy (CAP) framework, the segmentation results of each method were evaluated using the F-score and intersection over union (IoU) metrics. The respective average F-score and IoU obtained by each method are 0.682 and 0.524 for Mask R-CNN, 0.781 and 0.646 for U-Net, 0.808 and 0.683 for FracTAL ResUNet, and 0.805 and 0.678 for the optimised MRS approach. This study shows that DNNs, particularly FracTAL ResUNet, can be effectively used for large-scale segmentation of agricultural fields from satellite images.
PubDate: 2023-06-07
DOI: 10.1007/s41064-023-00247-x
- VOX2BIM+ - A Fast and Robust Approach for Automated Indoor Point Cloud
Segmentation and Building Model Generation-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Building Information Modeling (BIM) plays a key role in digital design and construction and promises also great potential for facility management. In practice, however, for existing buildings there are often either no digital models or existing planning data is not up-to-date enough for use as as-is models in operation. While reality-capturing methods like laser scanning have become more affordable and fast in recent years, the digital reconstruction of existing buildings from 3D point cloud data is still characterized by much manual work, thus giving partially or fully automated reconstruction methods a key role. This article presents a combination of methods that subdivide point clouds into separate building storeys and rooms, while additionally generating a BIM representation of the building’s wall geometries for use in CAFM applications. The implemented storeys-wise segmentation relies on planar cuts, with candidate planes estimated from a voxelized point cloud representation before refining them using the underlying point data. Similarly, the presented room segmentation uses morphological operators on the voxelized point cloud to extract room boundaries. Unlike the aforementioned spatial segmentation methods, the presented parametric reconstruction step estimates volumetric walls. Reconstructed objects and spatial relations are modelled BIM-ready as IFC in one final step. The presented methods use voxel grids to provide relatively high speed and refine their results by using the original point cloud data for increased accuracy. Robustness has proven to be rather high, with occlusions, noise and point density variations being well-tolerated, meaning that each method can be applied to data acquired with a variety of capturing methods. All approaches work on unordered point clouds, with no additional data being required. In combination, these methods comprise a complete workflow with each singular component suitable for use in numerous scenarios.
PubDate: 2023-05-30
DOI: 10.1007/s41064-023-00243-1
- Automation Strategies for the Photogrammetric Reconstruction of Pipelines
-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract A responsible use of energy resources is currently more important than ever. For the effective insulation of industrial plants, a three-camera measurement system was, therefore, developed. With this system, the as-built geometry of pipelines can be captured, which is the basis for the production of a precisely fitting and effective insulation. In addition, the digital twin can also be used for Building Information Modelling, e.g. for planning purposes or maintenance work. In contrast to the classical approach of processing the images by calculating a point cloud, the reconstruction is performed directly on the basis of the object edges in the image. For the optimisation of the, initially purely geometrically calculated components, an adjustment approach is used. In addition to the image information, this approach takes into account standardised parameters (such as the diameter) as well as the positional relationships between the components and thus eliminates discontinuities at the transitions. Furthermore, different automation approaches were developed to facilitate the evaluation of the images and the manual object recognition in the images for the user. For straight pipes, the selection of the object edges in one image is sufficient in most cases to calculate the 3D cylinder. Based on the normalised diameter, the missing depth can be derived approximately. Elbows can be localised on the basis of coplanar neighbouring elements. The other elbow parameters can be determined by matching the back projection with the image edges. The same applies to flanges. For merging multiple viewpoints, a transformation approach is used which works with homologous components instead of control points and minimises the orthogonal distances between the component axes in the datasets.
PubDate: 2023-05-22
DOI: 10.1007/s41064-023-00244-0
- Editorial
-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
PubDate: 2023-05-11
DOI: 10.1007/s41064-023-00241-3
- Semantic Real-Time Mapping with UAVs
-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Whilst mapping with UAVs has become an established tool for geodata acquisition in many domains, certain time-critical applications, such as crisis and disaster response, demand fast geodata processing pipelines rather than photogrammetric post-processing approaches. Based on our 3D-capable real-time mapping pipeline, this contribution presents not only an array of optimisations of the original implementation but also an extension towards understanding the image content with respect to land cover and object detection using machine learning. This paper (1) describes the pipeline in its entirety, (2) compares the performance of the semantic labelling and object detection models quantitatively and (3) showcases real-world experiments with qualitative evaluations.
PubDate: 2023-05-11
DOI: 10.1007/s41064-023-00242-2
- Reports
-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
PubDate: 2023-04-14
DOI: 10.1007/s41064-023-00236-0
- A Globally Applicable Method for NDVI Estimation from Sentinel-1 SAR
Backscatter Using a Deep Neural Network and the SEN12TP Dataset-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Vegetation monitoring is important for many applications, e.g., agriculture, food security, or forestry. Optical data from space-borne sensors and spectral indices derived from their data like the normalised difference vegetation index (NDVI) are frequently used in this context because of their simple derivation and interpretation. However, optical sensors have one major drawback: cloud coverage hinders data acquisition, which is especially troublesome for moderate and tropical regions. One solution to this problem is the use of cloud-penetrating synthetic aperture radar (SAR) sensors. Yet, with very different image characteristics of optical and SAR data, an optical sensor cannot be easily replaced by SAR sensors. This paper presents a globally applicable model for the estimation of NDVI values from Sentinel-1 C-band SAR backscatter data. First, the newly created dataset SEN12TP consisting of Sentinel-1 and -2 images is introduced. Its main features are the sophisticated global sampling strategy and that the images of the two sensors are time-paired. Using this dataset, a deep learning model is trained to regress SAR backscatter data to NDVI values. The benefit of auxiliary input information, e.g., digital elevation models, or land-cover maps is evaluated experimentally. After selection of the best model configuration, another experimental evaluation on a carefully selected hold-out test set confirms that high performance, low error, and good level of spatial detail are achieved. Finally, the potential of our approach to create dense NDVI time series of frequently clouded areas is shown. One limit of our approach is the neglect of the temporal characteristics of the SAR and NDVI data, since only data from a single date are used for prediction.
PubDate: 2023-04-13
DOI: 10.1007/s41064-023-00238-y
- Uncovering Early Traces of Bark Beetle Induced Forest Stress via
Semantically Enriched Sentinel-2 Data and Spectral Indices-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Forest ecosystems are shaped by both abiotic and biotic disturbances. Unlike sudden disturbance agents, such as wind, avalanches and fire, bark beetle infestation progresses gradually. By the time infestation is observable by the human eye, trees are already in the final stages of infestation—the red- and grey-attack. In the relevant phase—the green-attack—biochemical and biophysical processes take place, which, however, are not or hardly visible. In this study, we applied a time series analysis based on semantically enriched Sentinel-2 data and spectral vegetation indices (SVIs) to detect early traces of bark beetle infestation in the Berchtesgaden National Park, Germany. Our approach used a stratified and hierarchical hybrid remote sensing image understanding system for pre-selecting candidate pixels, followed by the use of SVIs to confirm or refute the initial selection, heading towards a 'convergence of evidence approach’. Our results revealed that the near-infrared (NIR) and short-wave-infrared (SWIR) parts of the electromagnetic spectrum provided the best separability between pixels classified as healthy and early infested. Referring to vegetation indices, we found that those related to water stress have proven to be most sensitive. Compared to a SVI-only model that did not incorporate the concept of candidate pixels, our approach achieved distinctively higher producer’s accuracy (76% vs. 63%) and user’s accuracy (61% vs. 42%). The temporal accuracy of our method depends on the availability of satellite data and varies up to 3 weeks before or after the first ground-based detection in the field. Nonetheless, our method offers valuable early detection capabilities that can aid in implementing timely interventions to address bark beetle infestations in the early stage.
PubDate: 2023-04-13
DOI: 10.1007/s41064-023-00240-4
- Automatic Detection of Specific Constructions on a Large Scale Using Deep
Learning in Very High Resolution Airborne Imagery-
Free pre-print version: Loading...
Rate this result:
What is this?
Please help us test our new pre-print finding feature by giving the pre-print link a rating.
A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract In the High Modernism period, from around 1914 to 1970, many system halls in steel construction were manufactured to meet the increasing demand in industry, commerce, and agriculture, among other areas. However, these types of buildings have not been the focus of any research in the field of construction history, generating a lack of knowledge regarding their construction types, distribution, and related context to enable statements on the ability and worthiness of historical monument listings. This paper proposes a methodology for the automatic detection of these buildings using aerial imagery. For this purpose, Deep Learning techniques for two tasks are evaluated: semantic segmentation and object detection. Different state-of-the-art software architectures are extensively reviewed and assessed through a series of experiments to determine which features and hyper-parameters produce the best results. Based on our experiments, the height information from nDSM improved the results by refining the detections and reducing the number of false negatives and false positives. Moreover, the Focal Loss helped boost the detections by tuning its hyper-parameter \(\gamma\) , where object detection algorithms showed high sensitivity to this value. Semantic segmentation models outperformed their counterparts for object detection, with U-Net and EfficientNet B3 as the backbone, the one with the best results with a \(Detection\ Rate\) of up to \(93\%\) .
PubDate: 2023-04-06
DOI: 10.1007/s41064-023-00237-z
|
Hybrid journal (It can contain Open Access articles)
