A New Maximum-Likelihood Joint Segmentation Technique for Multitemporal SAR and

In this paper, we devise a new technique for the fusion of a sequence of multitemporal single-channel synthetic aperture radar (SAR) images of a given area with a single multiband optical image. Unlike for SAR, the availability of optical images is largely affected by atmospheric conditions, so that this is a case of practical interest. First, a statistical model for the joint distribution of SAR and optical data is provided. Then, a split-merge test based on this model is derived, and its performance is evaluated both analytically and using a Monte Carlo simulation. A new segmentation technique is introduced (OPT MUM), based on the test and on a region-growing scheme. The effectiveness of the proposed technique for the fusion of multitemporal SAR and multiband optical images is tested on synthetic and real images. Results show that the proposed scheme allows to both 1) discriminate characteristics that would be impossible to distinguish using only a single sensor and 2) increase the overall discrimination performance, even when each sensor has its own discrimination capability.