Heterogeneous SAR texture characterization by means of Markov random fields

The purpose of this paper is to develop a new method for heterogeneous SAR texture modeling and characterization. It reposed on a Markov random field modeling. The K-distribution, derived under multiplicative model assumption, is shown to be an efficient model to characterize SAR signals at global scale. In order to characterize details of the texture, a local Markov model "auto-gamma" is derived and based on the global model.