SUPERVISED CLASSIFICATION OF K-DISTRIBUTED SAR IMAGES OF NATURAL TARGETS AND PROBABILITY OF ERROR ESTIMATION

A radiometric and textural classification method for the single-channe l synthetic: aperture radar (SAR) image is proposed, which explicitly takes into account the probability density function (pdf) of the under lying cross section for K-distributed images, This method makes extens ive use of adaptive preprocessing methods (e.g., Gamma-Gamma MAP speck le filtering), in order to ensure good classification accuracy as well as fair preservation of the spatial resolution of the final result, E rror rates can be estimated during the training step, allowing one to select only relevant reflectivity classes and to save computation time in trials, The classification method is based on a Maximum Likelihood (ML) segmentation of the filtered image, Finally, a texture criterion is introduced to improve the accuracy of the classification result.