Classification of Segments in PolSAR Imagery by Minimum Stochastic Distances Between Wishart Distributions

Abstract

A new classifier for Polarimetric SAR (PolSAR) images is proposed and assessed in this paper. Its input consists of segments, and each one is assigned the class which minimizes a stochastic distance. Assuming the complex Wishart model, several stochastic distances are obtained from the h-φ family of divergences, and they are employed to derive hypothesis test statistics that are also used in the classification process. This article also presents, as a novelty, analytic expressions for the test statistics based on the following stochastic distances between complex Wishart models: Kullback-Leibler, Bhattacharyya, Hellinger, Rényi, and Chi-Square; also, the test statistic based on the Bhattacharyya distance between multivariate Gaussian distributions is presented. The classifier performance is evaluated using simulated and real PolSAR data. The simulated data are based on the complex Wishart model, aiming at the analysis of the proposal well controlled data. The real data refer to the complex L-band image, acquired during the 1994 SIR-C mission. The results of the proposed classifier are compared with those obtained by a Wishart per-pixel/contextual classifier, and we show the better performance of the region-based classification. The influence of the statistical modeling is assessed by comparing the results using the Bhattacharyya distance between multivariate Gaussian distributions for amplitude data. The results with simulated data indicate that the proposed classification method has a very good performance when the data follow the Wishart model. The proposed classifier also performs better than the per-pixel/contextual classifier and the Bhattacharyya Gaussian distance using SIR-C PolSAR data.

Optical images-based edge detection in Synthetic Aperture Radar images

“…The methods were applied on twenty samples of a scene, which were simulated using Wishart distributions derived from a fully polarimetric image [13]. Using both visual inspection and the Baddeley Delta metric [14] we verified that the combination with the Lopez-Molina technique with the 9 Â 9 neighbourhood proposed by Fu et al [10] and preprocessing with the Enhanced Lee filter produced the best results.…”

“…We apply the methods on data derived from a fully polarimetric image, presented by da Silva et al [13], from an agricultural area in the Amazon region in Brazil (see Fig. 4).…”

“…5 illustrates the approach. For our study we apply the edge detection methods on twenty independently simulated mosaics amplitude images, using the parameters estimated in [13] to assess the quality of the methods.…”

“…da Silva et al [13] describe a classification experiment, based on a full polarimetric image from an agricultural area in the Amazon region in Brazil. In that study, the authors estimated the parameters for probability distributions associated to each of the classes of interest, such as water and different types of vegetation and their phenology.…”

“…They assessed their results in an image formed by a mosaic of the classes, with pixel values generated using the parameters found for each class. We apply all techniques addressed in this study on twenty simulated mosaics, using the parameters estimated in [13], considering amplitude images derived from different polarizations. We assess the quality of the results, according to Baddeley's Delta Metric (BDM) [14].…”