The /spl Gscr//sup 0//sub A/ distribution as the true model for SAR images

For many synthetic aperture radar (SAR) sensors, urban areas, forests and pasture areas have a growing degree of homogeneity, and for each one of these region types there is a distribution class that fits the data best. In this paper, the feasibility of using the /spl Gscr//sup 0//sub A/ distribution as a general model is studied, and the practical usefulness of this proposal is shown. The /spl Gscr//sup 0//sub A/ distribution is adopted as the true model, and its locally estimated parameters are used for classification of synthetic and real SAR images.