Authors:
Wen; J.;Tian, Z.;Liu, X.;Lin, W.; Pages: 759 - 768 Abstract: In this paper, we propose a manifold geometry based projective nonnegative matrix factorization linear dimensionality reduction method, called neighborhood preserving orthogonal projective nonnegative matrix factorization (NPOPNMF), for feature extraction of hyperspectral image. By adding constraints on projective nonnegative matrix factorization (PNMF) that each data point can be represented as a linear combination of its neighbors, NPOPNMF preserves local neighborhood geometrical structure of hyperspectral data in the reduced space, and overcomes the Euclidean limitation of PNMF. The metric structure of original high-dimensional hyperspectral data space is preserved due to the orthogonality of projection matrix. NPOPNMF can be performed in either supervised or unsupervised mode according to the construction of adjacency graph and it can improve the discriminant performance of PNMF. Theoretical analysis and experimental results on hyperspectral data sets demonstrate that the proposed method is an effective and promising method for hyperspectral image feature extraction. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Khazai; S.;Safari, A.;Mojaradi, B.;Homayouni, S.; Pages: 769 - 778 Abstract: Fast detecting difficult targets such as subpixel objects is a fundamental challenge for anomaly detection (AD) in hyperspectral images. In an attempt to solve this problem, this paper presents a novel but simple approach based on selecting a single feature for which the anomaly value is the maximum. The proposed approach applied in the original feature space has been evaluated and compared with relevant state-of-the-art AD methods on Target Detection Blind Test data sets. Preliminary results suggest that the proposed method can achieve better detection performance than its counterparts. The results also show that the proposed method is computationally expedient. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Mahmood; Z.;Akhter, M.A.;Thoonen, G.;Scheunders, P.; Pages: 779 - 791 Abstract: This paper describes a hyperspectral image classification method to obtain classification maps at a finer resolution than the image's original resolution. We assume that a complementary color image of high spatial resolution is available. The proposed methodology consists of a soft classification procedure to obtain landcover fractions, followed by a subpixel mapping of these fractions. While the main contribution of this article is in fact the complete multisource framework for obtaining a subpixel map, the major novelty of this subpixel mapping approach is the inclusion of contextual information, obtained from the color image. Experiments, conducted on two hyperspectral images and one real multi source data set, show excellent results, when compared to classification of the hyperspectral data only. The advantage of the contextual approach, compared to conventional subpixel mapping approaches, is clearly demonstrated. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Samiappan; S.;Prasad, S.;Bruce, L.M.; Pages: 792 - 800 Abstract: Traditional statistical classification approaches often fail to yield adequate results with Hyperspectral imagery (HSI) because of the high dimensional nature of the data, multimodal class distribution and limited ground truth samples for training. Over the last decade, Support Vector Machines (SVMs) and Multi-Classifier Systems (MCS) have become popular tools for HSI analysis. Random Feature Selection (RFS) for MCS is a popular approach to produce higher classification accuracies. In this study, we present a Non-Uniform Random Feature Selection (NU-RFS) within a MCS framework using SVM as the base classifier. We propose a method to fuse the output of individual classifiers using scores derived from kernel density estimation. This study demonstrates the improvement in classification accuracies by comparing the proposed approach to conventional analysis algorithms and by assessing the sensitivity of the proposed approach to the number of training samples. These results are compared with that of uniform RFS and regular SVM classifiers. We demonstrate the superiority of Non-Uniform based RFS system with respect to overall accuracy, user accuracies, producer accuracies and sensitivity to number of training samples. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Molero; J.M.;Garzon, E.M.;Garcia, I.;Plaza, A.; Pages: 801 - 814 Abstract: Anomaly detection is an important task for hyperspectral data exploitation. A standard approach for anomaly detection in the literature is the method developed by Reed and Xiaoli, also called RX algorithm. A variation of this algorithm consists of applying the same concept to a local sliding window centered around each image pixel. The computational cost is very high for RX algorithm and it strongly increases for its local versions. However, current advances in high performance computing help to reduce the run-time of these algorithms. So, for the standard RX, it is possible to achieve a processing time similar to the data acquisition time and to increase the practical interest for its local versions. In this paper, we discuss several optimizations which exploit different forms of acceleration for these algorithms. First, we explain how the calculation of the correlation matrix and its inverse can be accelerated through optimization techniques based on the properties of these particular matrices and the efficient use of linear algebra libraries. Second, we describe parallel implementations of the RX algorithm, optimized for multicore platforms. These are well-known, inexpensive and widely available high performance computing platforms. The ability to detect anomalies of the global and local versions of RX is explored using a wide set of experiments, using both synthetic and real data, which are used for comparing the optimized versions of the global and local RX algorithms in terms of anomaly detection accuracy and computational efficiency. The synthetic images have been generated under different noise conditions and anomalous features. The two real scenes used in the experiments are a hyperspectral data set collected by NASA's Airborne Visible Infra-Red Imaging Spectrometer (AVIRIS) system over the World Trade Center (WTC) in New York, five days after the terrorist attacks, and another data set collected by the HYperspectral Digital Image Collection Experiment (H-
DICE). Experimental results indicate that the proposed optimizations can significantly improve the performance of the considered algorithms without reducing their anomaly detection accuracy. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Wu; C.;Du, B.;Zhang, L.; Pages: 815 - 830 Abstract: Remote sensing change detection has played an important role in many applications. Most traditional change detection methods deal with single-band or multispectral remote sensing images. Hyperspectral remote sensing images offer more detailed information on spectral changes so as to present promising change detection performance. The challenge is how to take advantage of the spectral information at such a high dimension. In this paper, we propose a subspace-based change detection (SCD) method for hyperspectral images. Instead of dealing with band-wise changes, the proposed method measures spectral changes. SCD regards the observed pixel in the image of Time 2 as target and constructs the background subspace using the corresponding pixel in the image of Time 1, and additional information. In this paper, two types of additional information, i.e., spatial information in the neighborhood of the corresponding pixel in Time 1, and the spectral information of undesired land-cover types, are used to construct the background subspace for special applications. The subspace distance is calculated to determine whether the target is anomalous with respect to the background subspace. The anomalous pixels are considered as changes. Here, orthogonal subspace projection is employed to calculate the subspace distance, which makes full use of the advantage of the abundant spectral information in hyperspectral imagery, and is also easy to apply. The experimental results using Hyperion data and HJ-1A HSI data indicate that SCD gives more accurate detection results, with a lower false alarm rate, compared with other state-of-the-art methods. SCD with additional information also gives satisfactory results in the experiments, reducing the false alarms caused by misregistration and suppressing the change of undesired land-cover types. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Niazmardi; S.;Homayouni, S.;Safari, A.; Pages: 831 - 839 Abstract: Unsupervised classification approaches, also known as “clustering algorithms”, can be considered a solution to problems associated with the supervised classification of remotely sensed image data. The most important of these problems with respect to statistical classification algorithms is the lack of enough high quality training data and high dimensionality of hyperspectral data. In this paper, an improved clustering framework is developed and evaluated as a resolution to these problems. The proposed method enhances the Fuzzy C-Means (FCM) algorithm by using the Support Vector Domain Description (SVDD). The proposed algorithm operates in a similar manner as the FCM for the clustering and labeling of data vectors. However, for estimation of the cluster centers, the SVDD encircles the corresponding members and estimates the center of a containing sphere. By doing so, the effects of noise and outliers on the cluster centers are reduced, and more specifically, higher classification accuracy can be obtained. In spite of this advantage, there are two sets of parameters, namely, the SVDD's and FCM's parameters, both of which affect the performance of the proposed algorithm. Accordingly, the effects of these parameters and their optimum values have been evaluated as well. The evaluations of the results of experiments show that the proposed algorithm, due to the use of the SVDD algorithm, is more efficient than other clustering algorithms. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Jiang; B.;Liang, S.;Townshend, J.R.;Dodson, Z.M.; Pages: 840 - 850 Abstract: Data from the Chinese Huan-Jin (which means “environment”) 1 satellites, HJ-1A and HJ-1B, have been widely used for environmental, disaster monitoring and other applications. However, the radiometric properties of their CCD sensors have not been well assessed. In this study, we evaluated the radiometric performance of the HJ-1A/B CCD sensors by comparing their top-of-atmosphere (TOA) reflectance with those of three other satellite sensors – the Landsat-5 Thematic Mapper (TM) because of its long-term stable radiometric calibration, the Earth Observer-1(EO-1) Advanced Land Imager (ALI), and the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) – in four bands from the visible to near infrared spectrum over different landscapes. The results demonstrate that the radiometric performance of the HJ-1A/B CCD sensors is close to that of the Landsat-5 TM, ALI, and ASTER sensors, with the mean R
$^{2}$ values ranging from 0.95 to 0.96 (TM), 0.92–0.95 (ALI) and 0.87–0.93 (ASTER), and the mean normalized root mean square error values from 0.05–0.08 (TM), 0.05–0.09 (ALI) and 0.05–0.06 (ASTER). We further assessed the quality of HJ-1 CCD data using four image statistics and found that the image quality is very similar to that of the Landsat-5 TM. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Uto; K.;Seki, H.;Saito, G.;Kosugi, Y.; Pages: 851 - 860 Abstract: A low-cost, small, lightweight hyperspectral sensor system that can be loaded onto small unmanned autonomous vehicle (UAV) platforms has been developed for the acquisition of aerial hyperspectral data. Safe and easy observation is possible under unstable illumination conditions by using lightweight and autonomous cruising. The hyperspectral sensor system, equipped with a 256-band hyperspectral sensor covering a spectral range from 340–763 nm, a GPS and a data logger, is 400 g in total weight. The acquisition period for each sampling, 768 bytes, is 100 ms. The aerial hyperspectral data of rice paddies are collected under cloudy weather. The flight altitude from the ground is 10 m, and the cruising speed is 2 m/s. The high-accuracy estimation of the chlorophyll densities is confirmed, even under unstable illumination conditions, by frequent monitoring of the illumination level and the chlorophyll indices, based on the red-edge (RE) and near infrared (NIR) spectral ranges. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Senthilnath, J.;Omkar, S.N.;Mani, V.;Karnwal, N.;P.B; S.; Pages: 861 - 866 Abstract: The presence of a large number of spectral bands in the hyperspectral images increases the capability to distinguish between various physical structures. However, they suffer from the high dimensionality of the data. Hence, the processing of hyperspectral images is applied in two stages: dimensionality reduction and unsupervised classification techniques. The high dimensionality of the data has been reduced with the help of Principal Component Analysis (PCA). The selected dimensions are classified using Niche Hierarchical Artificial Immune System (NHAIS). The NHAIS combines the splitting method to search for the optimal cluster centers using niching procedure and the merging method is used to group the data points based on majority voting. Results are presented for two hyperspectral images namely EO-1 Hyperion image and Indian pines image. A performance comparison of this proposed hierarchical clustering algorithm with the earlier three unsupervised algorithms is presented. From the results obtained, we deduce that the NHAIS is efficient. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Verrelst; J.;Alonso, L.;Rivera Caicedo, J.P.;Moreno, J.;Camps-Valls, G.; Pages: 867 - 874 Abstract: Precise and spatially-explicit knowledge of leaf chlorophyll content
$(Chl)$
is crucial to adequately interpret the chlorophyll fluorescence $(ChF)$
signal from space. Accompanying information about the reliability of the
$Chl$
estimation becomes more important than ever. Recently, a new statistical method was proposed within the family of nonparametric Bayesian statistics, namely Gaussian Processes regression (GPR). GPR is simpler and more robust than their machine learning family members while maintaining very good numerical performance and stability. Other features include: i) GPR requires a relatively small training data set and can adopt very flexible kernels, ii) GPR identifies the relevant bands and observations in establishing relationships with a variable, and finally iii) along with pixelwise estimations GPR provides accompanying confidence intervals. We used GPR to retrieve
$Chl$ from hyperspectral reflectance data and evaluated the portability of the regression model to other images. Based on field $Chl$
measurements from the SPARC dataset and corresponding spaceborne CHRIS spectra (acquired in 2003, Barrax, Spain), GPR developed a regression model that was excellently validated (
$r^{2}$
: 0.96, RMSE: 3.82
$mu{rm g/cm}^{2}$ ). The SPARC-trained GPR model was subsequently applied to CHRIS images (Barrax, 2003, 2009) and airborne CASI flightlines (Barrax 2009) to generate
$Chl$
maps. The accompanying confidence maps provided insight -
n the robustness of the retrievals. Similar confidences were achieved by both sensors, which is encouraging for upscaling
$Chl$ estimates from field to landscape scale. Because of its robustness and ability to deliver confidence intervals, GPR is evaluated as a promising candidate for implementation into
$ChF$ processing chains. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Ni; W.;Sun, G.;Guo, Z.;Zhang, Z.;He, Y.;Huang, W.; Pages: 875 - 886 Abstract: Mapping of forest biomass over large area and in higher accuracy becomes more and more important for researches on global carbon cycle and climate change. The feasibility and problems of forest biomass estimations based on lookup table (LUT) methods using ALOS PALSAR data are investigated in this study. Using of the forest structures from a forest growth model as inputs to a three dimensional radar backscattering model, a lookup table is built. Two types of searching methods (Nearest Distance (ND) and Distance Threshold (DT)) are used to find solutions from lookup table. When a simulated dataset is used to test the lookup table, the RMSE of biomass estimation are 39.133 Mg/ha (R
$^{2}$
= 0.748) from ND and 26.699 Mg/ha (R
$^{2}$ = 0.886) from DT using dual-polarization data for forest with medium rough soil surface. All results show that DT is superior to ND. Comparisons of biomass from forest inventory data with that inversed from look up table using DT method over eight forest farms shows RMSE of 18.564 Mg/ha and 15.392 Mg/ha from PALSAR data with and without correction of the scattering mechanism, respectively. For the entire Lushuihe forest Bureau, the errors of the biomass estimation are
$-$ 13.8 Mg/ha (
${-}$ 8.6%) and
$-$ 5.5 Mg/ha (
$-$ 3.5%) using PALSAR data with and without correction of scattering mechanisms due to terrain, respectively. The results shows that the radar image corrected data could be directly used for biomass estimation using the lookup table method. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Liu; P.-W.;De Roo, R.D.;England, A.W.;Judge, J.; Pages: 887 - 899 Abstract: The performances of the soil moisture retrieval and assimilation algorithms using microwave observations rely on realistic estimates of brightness temperatures
$({rm T}_{rm B})$
from microwave emission models. This study identifies circumstances when current models fail to reliably relate near-surface soil moisture to an observed
${rm T}_{rm B}$
at L-band; offers a plausible explanation of the physical cause of these failures; and recommends improvements needed so that L-band observations can provide reliable estimates of soil moisture, more universally. Physically consistent soil parameters and moisture at the surface were estimated by using dual-polarized C-band observations during an intensive field experiment, for an irrigation event and subsequent drydown. These derived parameters were used in conjunction with the in situ moisture in deeper layers and different moisture profiles within the moisture sensing depth to obtain estimates of H-pol
${rm T}_{rm B}$
at L-band, that provided best matches with the observed
${rm T}_{rm B}$ . The general assumptions of linear moisture distribution, with uniform or exponentially decaying weighting functions provided realistic
${rm T}_{rm B}$
during the later stages of the drydown. However, the RMSDs of the ${rm T}_{rm B}{rm s}$
were upto 10.37 K during the wet period. In addition, the use of one value of moisture representing the entire moisture sensing depth during this highly dynamic stage of the drydown provides unrealistic estimates of emissivity, and hence,
${rm T}_{-
m B}$
at L-band. This study recommends use of a hydrological model to provide dynamic, realistic soil moisture profiles within the sensing depth and also an improved emissivity model that utilizes these detailed profiles for estimating
${rm T}_{rm B}$
. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Aubert; M.;Baghdadi, N.N.;Zribi, M.;Ose, K.;El Hajj, M.;Vaudour, E.;Gonzalez-Sosa, E.; Pages: 900 - 916 Abstract: TerraSAR-X data are processed for an “operational” mapping of bare soils moisture in agricultural areas. Empirical relationships between TerraSAR-X signal and soil moisture were established and validated over different North European agricultural study sites. The results show that the mean error on the soil moisture estimation is less than 4% regardless of the TerraSAR-X configuration (incidence angle, polarization) and the soil surface characteristics (soil surface roughness, soil composition). Furthermore, the potential of TerraSAR-X data (signal, texture features) to discriminate bare soils from other land cover classes in an agricultural watershed was evaluated. The mean signal backscattered from bare soils can be easily differentiated from signals from other land cover classes when the neighboring plots are covered by fully developed crops. This was observed regardless of the TerraSAR-X configuration and the soil moisture conditions. When neighboring plots are covered by early growth crops, a TerraSAR-X image acquired under wet conditions can be useful for discriminating bare soils. Bare soil masks were calculated by object-oriented classifications of mono-configuration TerraSAR-X data. The overall accuracies of the bare soils mapping were higher than 84% for validation based on object and pixel. The bare soils mapping method and the soil moisture relationships were applied to TerraSAR-X images to generate soil moisture maps. The results show that TerraSAR-X sensors provide useful data for monitoring the spatial variations of soil moisture at the within-plot scale. The methods of bare soils moisture mapping developed in this paper can be used in operational applications in agriculture, and hydrology. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Kumar; P.;Sharma, L.K.;Pandey, P.C.;Sinha, S.;Nathawat, M; Pages: 917 - 923 Abstract: This study focus on the biomass estimation of Sariska Wildlife Reserve using forest inventory and geospatial approaches to develop a model based on the statistical correlation between biomass measured at plot level and the associated spectral characteristics. The multistage statistical technique with incorporated the satellite data of IRS P-6 LISS III gives a precise estimation of biomass. Forest cover, forest stratum, and biomass maps were generated in the study. Spectral signatures along with tonal and textural variations were used to classify different forest types validated with GPS and ground truth data. Altitude dependent vegetation and contour information from toposheets were also considered while classifying imagery during interpretation. Sample plots were laid in study area with 0.1 ha area at intersect of the diagonals of the plots. DBH and height of all the trees inside the plot were measured and converted to biomass using volumetric equations depending upon specific gravity. The specific gravity of each tree species differ from each other and sometimes unique in different regions and varies from forest type of different regions. Estimation of tree biomass can serve as useful benchmark for future studies in related areas. Linear equation obtained was used as the model to generate final biomass map where predicted and estimated biomass were compared for each band of the satellite imageries. Linear, logarithm and power exponential models were compared to each other for correlation coefficient. Correlation between estimated and predicted AGB is 0.835 and coefficient of determination
$({rm r}^{2})$ value is 0.698. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Allouis; T.;Durrieu, S.;Vega, C.;Couteron, P.; Pages: 924 - 934 Abstract: The diameter at breast height (DBH) is the most extensively measured parameter in the field for estimating stem volume and aboveground biomass of individual trees. However, DBH can not be measured from airborne or spaceborne light detection and ranging (LiDAR) data. Consequently, volume and biomass must be estimated from LiDAR data using other tree metrics. The objective of this paper is to examine whether full-waveform (FW) LiDAR data can improve volume and biomass estimation of individual pine trees, when compared to usual discrete-return LiDAR data. Sets of metrics are derived from canopy height model (CHM-only metrics), from the vertical distribution of discrete-returns (CHM+DR metrics), and from full-waveform LiDAR data (CHM+FW metrics). In each set, the most relevant and non-collinear metrics were selected using a combination of methods using best subset and variance inflation factor, in order to produce predictive models of volume and biomass. CHM-only metrics (tree height and tree bounding volume [tree height x crown area] provided volume and biomass estimates of individual trees with an error (mean error $pm$
standard deviation) of 2%
$pm$
26% and
$-$
15% $pm $
49%, which is equivalent to previous studies. CHM+FW metrics did not improve stem volume estimates (5%
$pm$
31%), but they increased the accuracy of aboveground biomass estimates ( $-$
4%
$pm $
31%). The approach is limited by the delineation of individual trees. However, the results highlight the potential of full-waveform LiDAR data to improve-
aboveground biomass estimates through a better integration of branch and leaf biomass than with discrete-return LiDAR data. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Onojeghuo; A.O.;Blackburn, G.A.; Pages: 935 - 941 Abstract: Reedbeds are dominated by a small number of plant species, but are extremely valuable habitats for faunal biodiversity. However, reedbeds often exist in small patches distributed across landscapes and for most regions there is a lack of information about their location and condition. This paper investigates the potential of using LiDAR-derived elevation and intensity data to characterise reedbeds. A Leica ALS50 was used to acquire data for reedbeds during the leaf-off phenological period and the study site encompassed a wide range of canopy development. For reedbeds there was a lack of multiple LiDAR returns and ground returns, which limited the ability to acquire information on canopy structure or terrain elevation. Nevertheless, the first return LiDAR data was able to generate an accurate digital surface model and subsequent canopy height model, as validated using field measurements (RMSE 0.47 m; average difference 0.09 m (5% of average height)). LiDAR intensity data displayed specular reflection effects within reedbed areas, but off-nadir imagery was successfully used for mapping reedbeds, non-reedbed vegetation and water bodies. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Rao; W.;Li, G.;Wang, X.;Xia, X.-G.; Pages: 942 - 952 Abstract: It has been shown in the literature that, the inverse synthetic aperture radar (ISAR) echo can be seen as sparse and the ISAR imaging can be implemented by sparse recovery approaches. In this paper, we propose a new parametric weighted L
$_{1}$ minimization algorithm for ISAR imaging based on the parametric sparse representation of ISAR signals. Since the basis matrix used for sparse representation of ISAR signals is determined by the unknown rotation parameter of a moving target, we have to estimate both the ISAR image and basis matrix jointly. The proposed algorithm can adaptively refine the basis matrix to achieve the best sparse representation for the ISAR signals. Finally the high-resolution ISAR image is obtained by solving a weighted L $_{1}$
minimization problem. Both numerical and real experiments are implemented to show the effectiveness of the proposed algorithm. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Natsuaki; R.;Hirose, A.; Pages: 953 - 959 Abstract: Interferometric synthetic aperture radar (InSAR) is a useful technology to observe the earth topography. However, a synthetic aperture radar (SAR) interferogram usually includes a lot of rotational points, that is, singular points (SPs). SPs seriously affect the quality of generated digital elevation model (DEM). One of the dominant origins of the SPs is the local distortion in the co-registration of the master and slave images, which are the source of the interferogram. Previously, we proposed a local and fine co-registration method of the master and the slave using the number of SPs as the evaluation criterion (SPEC method). In this paper, we propose an improved version of the SPEC method which uses the shape-from-shading technique additionally for a better adjustment. In comparison to the conventional SPEC method, results indicate that the proposed technique improves the signal-to-noise ratio of the created DEM from InSAR images. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Lombardini; F.;Cai, F.;Pasculli, D.; Pages: 960 - 968 Abstract: Synthetic Aperture Radar Tomography (Tomo-SAR) is an emerging experimental “coherent data combination” mode allowing unprecedented full 3-D imaging of complex urban and infrastructure scenarios with layover (“garbled”) scatterers, exploiting multibaseline interferometric SAR data stacks. Various approaches have been proposed to improve Fourier-based Tomo-SAR elevation beamforming which is affected by unsatisfactory height sidelobe behaviour and resolution, due to the typical low number of baselines with irregular distribution. Among these approaches, height superresolution multilook beamforming techniques proved to posses interesting capabilities, at the cost of operation with reduced horizontal resolution. In this work, a recently proposed knowledge-based baseline interpolation and the Capon and MUSIC superresolution methods are integrated in to a new Tomo-SAR processor able to offer at a low computational burden height superresolution and sidelobe cleaning with single-look data, allowing full resolution operation, as important in urban and other man-made areas. Results are reported with real ERS data. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Makarau; A.;Palubinskas, G.;Reinartz, P.; Pages: 969 - 990 Abstract: The way of multisensory data integration is a crucial step of any data fusion method. Different physical types of sensors (optic, thermal, acoustic, or radar) with different resolutions, and different types of GIS digital data (elevation, vector map) require a proper method for data integration. Incommensurability of the data may not allow to use conventional statistical methods for fusion and processing of the data. A correct and established way of multisensory data integration is required to deal with such incommensurable data as the employment of an inappropriate methodology may lead to errors in the fusion process. To perform a proper multisensory data fusion several strategies were developed (Bayesian, linear (log linear) opinion pool, neural networks, fuzzy logic approaches). Employment of these approaches is motivated by weighted consensus theory, which lead to fusion processes that are correctly performed for the variety of data properties. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Gong; X.;Corpetti, T.; Pages: 991 - 1003 Abstract: Many problems related to change detection require to compute image features on local windows. Such features usually combine in each pixel locations spectral values (luminance) associated with some spatial properties, such as texture features or more advanced local relationships between pixels. Therefore, as far as local windows are considered, the optimal size selection is a key point for the performance of the algorithm. This paper tackles this issue by proposing an original mean to estimate the size of local windows at each pixel. It uses a stochastic representation of the image grid. By combining some rules of stochastic calculus, we redefine image features functions on specific image grids where the changed areas are modeled. This enables to extract in each location the optimal size on which image feature function should be consistent. For validation's sake, we propose a simple change detection approach that takes into account basic image features computed on local windows where the sizes are either manually fixed or automatically estimated, both using our approach and existing window size estimation techniques. In addition, some comparisons with state-of-the-art change detection methods are presented. This allows to validate the efficiency of our proposition and to demonstrate that even simple techniques associated with accurate image features can provide interesting results. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Miyazaki; H.;Shao, X.;Iwao, K.;Shibasaki, R.; Pages: 1004 - 1019 Abstract: We present an automated classification method for global urban area mapping by integrating satellite images taken by Visible and Near-Infrared Radiometer of Advanced Spaceborne Thermal Emission and Reflection radiometer (ASTER/VNIR) and GIS data derived from existing urban area maps. The method consists of two steps. First, we extracted urban areas from ASTER/VNIR satellite images by using an iterative machine-learning classification method known as Learning with Local and Global Consistency (LLGC). This method is capable of automatically performing classification with a noisy training dataset, in our case, low-resolution urban maps. Therefore, we were able to perform supervised classification of ASTER/VNIR images without using labor-intensive visual interpretation. Second, we integrated the LLGC confidence map with other maps by logistic regression. The logistic regression complemented misclassifications in the LLGC map and provided useful information for further improvement of the model. In an experiment including 194 scenes of ASTER/VNIR images, the integrated maps were developed at a resolution of 15 m resolution, which is much finer than existing maps with resolutions of 300 to 1000 m. The maps achieved an overall accuracy of 90.0% and a kappa coefficient of 0.565, both of which are higher than or almost equal to the values for major existing global urban area maps. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Sheinker; A.;Ginzburg, B.;Salomonski, N.;Frumkis, L.;Kaplan, B.Z.; Pages: 1020 - 1030 Abstract: In this work we propose methods for object localization in 2D using beacons of low frequency quasi-static magnetic field. From a practical point of view, localization in 2D is sufficient for many applications, requiring much less calculations than in 3D, making it more robust and easier to implement in real-time low power applications. The low frequency magnetic field may penetrate foliage, soil, buildings, and many other types of media. This is an important advantage over traditional localization methods such as sonar or radar, where effective operation requires line-of-sight. Another advantage of the low frequency magnetic fields is that there is no direct influence by bad weather conditions and diurnal variations. Opposite to traditional electromagnetic methods, where operational range is usually more than a wavelength, low frequency induction approach results in a relatively limited localization range. Each beacon comprises a coil generating a magnetic field of a unique frequency in the ULF band. The generated magnetic fields are sensed by a search-coil magnetometer. The magnetometer readings are processed to estimate the magnitude and phase of the received beacons signals, which are used to localize the magnetometer. For a moving object, we propose to combine localization together with tracking algorithm using a data fusion approach. The proposed methods have been tested using numerous computer simulations, showing accurate localization results. A prototype was developed and used in field experiments, validating simulation results. The good accuracy together with a simple implementation makes the proposed methods attractive to many real-time low power field applications. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Jaiswal; N.;Kishtawal, C.M.; Pages: 1031 - 1035 Abstract: In the present work, an objective technique has been presented to fix the center position of TC in the satellite generated infrared images. The basis of the technique is to determine the point around which the fluxes of the gradient vectors of brightness temperature (BT) are converging. First, variance of brightness temperature at each pixel from its neighboring pixels is computed and then flux of the gradient of variance values is computed. Next, a line parallel to the gradient vector at each pixel is drawn across the image, and the locations where these lines intersect each other are stored in a density matrix. The score values accumulated in the density matrix are averaged and location with the highest score is identified. This position is considered to be the center location of the cyclone. The technique has been tested over the Kalpana satellite generated (approximately 1000) IR images of the cyclones that formed during the period 2009–2010. The technique has been used in fully automated mode for the four cyclones viz., Phyan, Ward, Laila, and Phet. The half hourly sequential IR images during the life period of each cyclone is analyzed and the center position is determined. The track of cyclone obtained by the automatically determined center position is compared with the observed track obtained from Joint Typhoon Warning Centre (JTWC). The mean track error with respect to JTWC observations for four cyclones Phyan, Ward, Laila, and Phet was computed and found as 42, 82, 58, and 42.5 km, respectively. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Sixian; Q.;Jianwen, M.;Xuanji, W.; Pages: 1036 - 1047 Abstract: A Hierarchical Bayesian Network Algorithm (HBN) is developed for data assimilation and tested with an instance of soil moisture assimilation from hydrological model and ground observations. In essence, HBN is a framework that can statistically describe Bayesian models and capture the dependences in the models more realistically than non-hierarchical Bayesian models. In this work, data assimilation separates into data level, process level and parameter level, and conditional probability models are defined for each level. The data model mainly deals with the scale differences between multiple data, while the process model is designed to take account of non-stationary process. Soil moisture from Soil Moisture Experiment in 2003 and Variable Infiltration Capacity Model is sequentially assimilated with HBN. The result shows that the assimilation with HBN provides spatial and temporal distribution information of soil moisture and the assimilation result agrees well with the ground observations. In summary, the HBN is a good algorithm together with data, process and parameter model, which shows great potential for data assimilation development. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
Authors:
Almendros-Jimenez; J.M.;Domene, L.;Piedra-Fernandez, J.A.; Pages: 1048 - 1063 Abstract: In this paper we present a framework for ocean image classification based on ontologies. With this aim, we will describe how low and high level content of ocean satellite images can be modeled with an ontology. In addition, we will show how the image classification can be modeled with the ontology in which decision tree based classifiers and rule-based expert systems are represented. Particularly, the rule based expert systems include rules about low-level features (called training and labeling rules), and rules defined from the labeling (called human expert rules). The modeling with the ontology provides an extensible framework in which accommodate several methods of image classification. One of the main aims of our proposal is to provide a mechanism to share data about image classification between applications. We have developed an extensible Protégé plugin to classify images. PubDate:
April 2013
Issue No:Vol. 6, No. 2 (2013)
LM Information Delivery
One of Europe's leading subscription and information management providers offering cost-efficient solutions for academic and research libraries.
SUNCAT is the largest freely available source of information about
serials holdings in the UK. Researchers are able to locate serials held
in 85 UK research libraries.
[17 followers] Follow 
Subscription journal
[163 journals]
@journaltocs
(4 followers) 
(4 followers) 
(4 followers) 
