On the use of a multispectral Markov random field model for texture analysis in multitemporal SAR imagery

In this paper we propose the use of the multispectral extension of the grey level Markov random field model for multitemporal texture analysis in synthetic aperture radar (SAR) imagery. This extension is not a straightforward one, due to problems specific to the nature of the multitemporal SAR images. To circumvent these problems, as well as the computational burden, we propose another simple method for the calculation of the model parameters. These models are examined for their texture characterisation ability, as well as their behaviour under several noise models.