Application of DSM theory for SAR image change detection

Synthetic Aperture Radar (SAR) data enables direct observation of land surface at repetitive intervals and therefore allows temporal detection and monitoring of land changes. However, the problem of radar automatic change detection is made more difficult, mainly with the presence of speckle noise. This paper presents a new method for SAR image change detection using the Dezert-Smarandache Theory (DSmT). First, a Gamma distribution function is used to characterize globally the radar texture data and allows mass assignment throw Kullback-Leibler distance. Then, local pixel measurements are introduced to refine the mass attribution and take into account the context information. Finally, DSmT is carried out by comparing the modelling results between temporal images. The originality of the proposed method is on the one hand, the use of DSmT which achieve a plausible and paradoxical reasoning comparing to classical Dempster-Shafer Theory (DST). On the other hand, the given approach characterizes the radar texture data with a Gamma distribution which allows a better representation of the speckle. The radar texture is being usually modeled by a Gaussian model in previous DST and DSmT fusion works.