Change detection for SAR images based on bivariate Gamma models

The traditional CFAR target detection algorithm is strongly restrained for the manmade targets immersed in the environment with strong scattered clutter. In order to improve the detection performance, this paper proposes a novel algorithm based on the bivariate Gamma distributions. In addition, some key steps such as parameter estimation, change analysis, CFAR normalization, and targets clustering are also discussed. This algorithm, based on high approximation accuracy of bivariate Gamma distributions, fully uses the correlation of images to suppress the strong scattered clutter. The results on actual data indicate this algorithm has a quite good detection performance and can realize a relatively high detection rate under the condition of a low false alarm rate.