Modelling of correlated Gamma-distributed texture based on spherically invariant random process

A Gamma-distributed texture in coherent imagery is widely applied on some fields such as description of texture characteristic and verification of classification algorithm. In this paper, a method based on spherically invariant random process for texture modeling in SAR imagery has been proposed, and the feasibility of modeling for Gamma-distributed texture with SIRP is testified, the representation functions have also been educed. By means of the relationship between the envelope and the complex SIRV, the correlation property has been introduced into the texture. And a modeling and simulating flow of correlated Gamma-distributed texture using SIRV theorem has been established. A computer simulation has been made for verification. In order to testify whether the results obey Gamma distribution, a maximum likelihood estimate fitting has been made. The results show: theoretical analysis fits the simulating data well, and the generated texture image is fairly realistic; using the proposed method, the image PDF and its ACF can be independently controlled, and the correlation property can be conveniently induced with no limitation on the form of ACF.