Abstract:
This paper presents a SAR image segmentation method that combines regular tessellation and the Metropolis-Hastings (M-H) algorithm. First the image domain is partitioned into a group of rectangular sub-blocks by regular tessellation and then the image is modeled on the assumption that intensities of its pixels in each homogeneous region follow an identical and independent Gamma distribution. A region-based SAR image segmentation model is built using the Bayesian paradigm. Then, an M-H scheme is used to simulate the segmentation model, which can segment SAR image and estimate the model parameters. In the M-H algorithm, three move types are designated, including updating parameter vector, updating label field, and splitting or merging sub-block. The results obtained from both real and simulated SAR images show that the proposed algorithm works effectively and efficiently.