CFAR Ship Detection in SAR Images Based on the Generalized Rayleigh Mixture Models

Abstract

Synthetic Aperture Radar (SAR) is a powerful equipment that has gained popularity as it synthetically produces higher resolution images in any weather conditions and even at night. For this reason, the SAR can be used to detect ships using Constant False Alarm Rate (CFAR) algorithms. In this paper, based on the Expectation-Maximization (EM) algorithm, we introduce the Generalized Rayleigh Mixture Model (GRMM), for characterizing sea clutter. In doing this, we use an adaptive global threshold to generate a censorship map that indicates if each sample in the image is likely a target pixel. Experiments carried out on a real SAR image, show that the GRMM-CFAR detector transcends the existing detectors.