Cédric Lardeux

Support Vector Machine for Multifrequency SAR Polarimetric Data Classification

The objective of this paper is twofold: first, to assess the potential of radar data for tropical vegetation cartography and, second, to evaluate the contribution of different polarimetric indicators that can be derived from a fully polarimetric data set. Because of its ability to take numerous and heterogeneous parameters into account, such as the various polarimetric indicators under consideration, a support vector machine (SVM) algorithm is used in the classification step. The contribution of the different polarimetric indicators is estimated through a greedy forward and backward method. Results have been assessed with AIRSAR polarimetric data polarimetric data acquired over a dense tropical environment. The results are compared to those obtained with the standard Wishart approach, for single frequency and multifrequency bands. It is shown that, when radar data do not satisfy the Wishart distribution, the SVM algorithm performs much better than the Wishart approach, when applied to an optimized set of polarimetric indicators.

Classification of Tropical Vegetation Using Multifrequency Partial SAR Polarimetry

This letter presents a case study addressing the comparison between different synthetic aperture radar (SAR) partial polarimetric options for tropical-vegetation cartography. These options include compact polarization (CP), dual polarization (DP), and alternating polarization (AP). They are all derived from fully polarimetric (FP) SAR data acquired by the airborne SAR (AIRSAR) sensor over the French Polynesian Tubuai Island. The classification approach is based on the support vector machine algorithm and is further validated by several ground surveys. For a single frequency band, FP data give significantly better results than any other partial polarimetric configuration. Among the partial polarimetric architectures, the CP mode performs best. In addition, the DP mode shows better performance than the AP mode, highlighting the value of the polarimetric differential phase. The combination of different frequency bands (P-, L-, and C-bands) holds the most significant improvement: The multifrequency diversity adds generally more information than the multipolarization diversity. A noticeable result is the major contribution of the C-band at VV polarization (the only polarization available at C-band with the AIRSAR data set used in this letter) to the classification performance, due to its ability to discriminate between Pinus and Falcata.

Use of the SVM Classification with Polarimetric SAR Data for Land Use Cartography

Yhis study comes within the framework of the global cartography and inventory of the Polynesian landscape. An AIRSAR airborne acquired fully polarimettric data in L and P bands, in August 2000, over the main Polynesian Islands. This study focuses on Tubuai Island, where several ground surveys allow the validation of the different results. Different decompositions, such as H/A/alpha , or based on the Pauli formalism have shown their potential for land use discrimination. In order to take into account these different parameters into a supervised classification scheme, the SVM (Support Vector Machine) method is investigated. When dealing with only the coherent matrix elements, the results show that the SVM classification gives comparative results to those obtain with Wishart classification. Results are significantly improved when adding to the coherent matrix elements, other polarimetric parameters, as H/A/alpha or the co-polarized circular polarization correlation coefficient, rhorrll, for the Support Vector definition. Finally the best results are given when merging all the parameters for P and L bands, in addition to the only VV single channel acquired in C band.

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING SYNERGY BETWEEN LiDAR, RADARSAT-2 AND SPOT-5 IMAGES FOR THE DETECTION AND MAPPING OF WETLAND VEGETATION IN THE DANUBE DELTA

Wetlands are among the most productive natural environments on Earth, as they harbor exceptional biological diversity. For this paper, our study site was the Danube Delta. The biodiversity of the Danube Delta is extraordinary and it possesses one of the largest reed beds in the world. The main goal of our paper was to recognize, characterize, and map the main vegetation units of the Danube Delta. The paper emphasizes the importance of the joint use of LiDAR measurements (acquired in May 2011), RADARSAT-2 radar data (acquired on June 4, 2011), and SPOT-5 optical data (acquired on May 25, 2011). LiDAR data allow for the characterization of vegetation height within centimeter accuracy (10 cm). The radar measurements are based on C-band, providing additional information about the structure of the vegetation cover. The simultaneous acquisition of HH, HV, VV, and VH polarizations enabled us to discriminate between the targets, depending on their responses to the various polarizations, by calculating their polarimetric signatures. By linking multispectral LiDAR and radar data, information can be obtained about vegetation reflectance and height as well as the backscattering mechanism, allowing for improved mapping and characterization accuracy (90.60% mean accuracy). An accuracy assessment of the classification results was evaluated against the vegetation data recorded in the field.

Change detection in floodable areas of the Danube delta using radar images

In the wetlands of the Danube delta floodplain, flooding is a major natural risk. The coastal wetlands have been seriously impacted by floods in 2002, 2005, 2006 and 2010. Using hydrological and satellite observations acquired in 2009 and during the summer of 2010, this paper tackles the issue of forecasting risk based on land cover information and observations. A major objective of this methodological work consists in exploring several types of data from the Japanese ALOS satellite. These data are used to illustrate a multi-temporal radar data processing methodology based on temporal entropy analysis that enables change detection in the floodable areas of the Danube delta.

Contribution of TerraSAR-X radar images texture for forest monitoring

This study aims to evaluate the texture analysis of high spatial resolution images for mapping tropical forests. More precisely, it evaluates the potential of TerraSAR-X image, with spatial resolution of 0.5 meter for the classification of tropical forests located in southern Cambodia. In particular, the focus is put on the contribution of the analysis of textural information for classification. This latter is apprehended through the analysis of Haralick textural parameters. The retained algorithm of classification is the Support Vector Machine, as it allows taking into account numerous parameters, which can be heterogeneous with respect to their physical dimension. First results show that the addition of Haralick parameters to intensity channel may improve significantly the accuracy of the classification results. However, their performance for classification discrimination strongly depends on the size of the neighborhood from which they are estimated. Preliminary analysis of variograms allows optimizing the choice of the neighborhood size. Best results are obtained with a 25×25 sliding window size, with a classification accuracy improvement higher than 50% is observed.

Radar Polar Decomposition for Natural Surfaces Cartography

Illustrations of parameters obtained from polar decomposition derived from fully polarimetric data acquired by ALOS-PALSAR over the Mai-Ndombo lake, in the Democratic Republic of Congo are presented. Results show their complementarity to usual polarimetric indices, such as the entropy or the alpha parameters

Alteration and Remediation of Coastal Wetland Ecosystems in the Danube Delta. A Remote-Sensing Approach

Wetlands are important and valuable ecosystems; yet, since 1900, more than 50% of wetlands have been lost worldwide. An example of altered and partially restored coastal wetlands is the Danube Delta in Romania. Over time, human intervention has manifested itself in more than one-quarter of the entire Danube surface. This intervention was brutal and has rendered ecosystem restoration very difficult. Studies for rehabilitation/re-vegetation were begun immediately after the Danube Delta was declared a Biosphere Reservation in 1990. Remote sensing offers accurate methods for detecting changes in restored wetlands. Vegetation change detection is a powerful indicator of restoration success. The restoration projects use vegetative cover as an important indicator of restoration success. To follow the evolution of the vegetation cover of the restored areas, images obtained by radar and optical satellites, such as Sentinel-1 and Sentinel-2, have been used. The sensitivity of such sensors to the landscape depends on the wavelength of the radar or optical detection system and, for radar data, on polarization. Combining these types of data, which are associated with the density and size of the vegetation, is particularly relevant for the classification of wetland vegetation. In addition, the high temporal acquisition frequencies used by Sentinel-1, which are not sensitive to cloud cover, allow the use of temporal signatures of different land covers. Thus, to better understand the signatures of the different study classes, we analyze the polarimetric and temporal signatures of Sentinel-1 data. In a second phase, we perform classifications based on the Random Forest supervised classification algorithm involving the entire Sentinel-1 time series, proceeding through a Sentinel-2 collection and finally involving combinations of Sentinel-1 and-2 data. The supervised classifier used is the Random Forest algorithm that is available in the OrfeoToolbox (version 5.6) free software. Random Forest is an ensemble learning technique that builds upon multiple decision trees and is particularly relevant when combining different types 2 of indicators. The results of this study relate to the use of combinations of data from different satellite sensors (multi-date Sentinel-1, Sentinel-2) to improve the accuracy of recognition and mapping of major vegetation classes in the restoring areas of the Danube Delta. First, the data from each sensor are classified and analyzed. The results obtained in the first step show quite good classification performance for only one Sentinel-2 data (87.5% mean accuracy), in contrast to the very good results obtained using the Sentinel-1 time series (95.7% mean accuracy). The combination of Sentinel-1 time series and optical data from Sentinel-2 improved the performance of the classification (97.1%).

Contribution of textural information from TerraSAR-X image for forest mapping

This study evaluates the potential of High Resolution Spotlight TerraSAR-X image for forest type discrimination. Emphasis is put on textural analysis accessible with high resolution radar data. Textural attributes are extracted from GLCM matrices, wavelet, and Fourier Transform (i.e. FOTO method). Their contribution for classification is assessed by their performance through the SVM algorithm.

Using texture from high resolution Terrasar-X images for tropical forest mapping

This study aims to evaluate the contribution of textural analysis from high spatial resolution TerraSAR-X images for tropical forests mapping. This study evaluates the potential of the High resolution Spotlight TerraSAR-X image (HS), with a spatial resolution of 0.6 meter in Range and 1.1 meter in Azimuth, for the classification of tropical forest plantation of native species. Indeed, the contribution of the analysis of textural information for classification has been emphasized through the analysis of Haralick textural parameters, a second order statistic parameters computed in a certain direction with a distance (d) and window size (w). The retained algorithm of classification is SVM (Support Vector Machine as it allows taking into account numerous parameters, which can be heterogeneous with respect to their physical dimension. To resolve the issue of class heterogeneity in the context of high resolution image, a post classification has been applied by the mean of a majority filter. In this case, the majority filter was weighted by integrating a shape of different plots containing the label of the tree planted species.