Supervised fusion-classification of multifrequency polarimetric SAR images using K distribution and theory of evidence

Importance of using multifrequency polarimetric SAR data for scene classification has been recognized in remote sensing applications. However, many problems have to be faced, due to the inadequacy of the conventional classification techniques because of the nature of the SAR images. This paper demonstrates the feasibility of using the theory of evidence to merge and classify this type of image. The data are modeled using the generalized K distribution. The suggested methodology is evaluating using SAR images provided by SIR-C.