We present a filtering framework based on a new statistical model for Single-Look Complex (SLC) Synthetic Aperture Radar (SAR) data, the Scaled Normal-Inverse Gaussian (SNIG). The real and imaginary parts of the SLC image are modeled as mixtures of SNIGs, and the clustering of the mixture components is conducted using a Stochastic Expectation-Maximization (SEM) algorithm. Model parameters are associated to each pixel according to its class, thus producing parametric images of the entire scene. A closed-form Maxmum A Posteriori (MAP) filter then delivers a de-speckled estimate of the image. The method is tested on RADARSAT-2 data, HV polarization, representing images of icebergs surrounded by ocean water off the coast of the Hopen Island (Svalbard archipelago). Post-processing, the iceberg Contrast-to-Noise Ratio (CNR) defined relative to the open water clutter is improved compared to Single-Look Intensity (SLI) image, increasing from a value of 11 to a value of 30 for one of the targets, and the Coefficient of Variation (CV) of the clutter is reduced to a fraction of 0.01 compared to the same reference. We conclude that the proposed method shows potential for improving iceberg detection in open water.
|