Bayesian restoration of high resolution SAR imagery with Gauss-Markov random fields

In this paper, we introduce a Bayesian restoration approach with Gauss-Markov random fields (GMRF) for high resolution SAR imagery. By adopting Bayesian analysis framework, the restoration model of degraded image of markov random field can be built, and then the problem of image restoration is transformed into the combined optimization problem of solving maximum a posterior (MAP) estimation of model or minimum energy function, random field model parameters can be also estimated directly from noise image, thus speckle is effectively reduced. A high-resolution airborne image is chosen for experiments, the results show that the proposed method outperforms standard local statistics adapted de-noising techniques in terms of speckle reducing and preservation of structural detail information.