An algorithm for SAR image segmentation

An algorithm for SAR image segmentation is proposed on the basis of features of SAR image and Bayesian detection theory. The algorithm consists of two steps: first, segmentation threshold is decided by using method proposed in this paper; second, blurred enhancement SAR image is segmented by using segmentation threshold acquired in one step to detect interesting objects, which is strongly reflecting objects. After lots of experiment, simulation results show that the algorithm can detect effectively strongly reflecting objects for SAR image, and restrain effectively clutter background.