Automated target recognition with SAS: Shadow and highlight-based classification

This paper presents an integrated technique for automated target recognition (ATR) in synthetic aperture sonar (SAS) images. The recognition procedure starts with a despeckling approach based on the anisotropic diffusion filter. As a second step, a fuzzymorpho-based segmentation procedure is applied to the filtered images, which partitions the image into highlights and shadow areas. A number of geometrical features are extracted from these areas, and are then used to classify targets using a Markov Chain Monte Carlo (MCMC) approach Very promising results are obtained.