New contributions into the Dezert-Smarandache theory: Application to remote sensing image classification

The theory of Dezert and Smarandache (DSmT) called also the theory of plausible and paradoxical reasoning represents a suitable formalism to fuse paradoxical sources. The application of the DSmT needs at first the generation of the hyper-powerset and more exactly the composites events. Thus, we present in this paper a new approach to construct pertinent paradoxical classes based on gray levels histograms, which allows also the reduction of the cardinality of the hyper-powerset. Then, we developed a new technique for ordering and coding generalized focal elements. This technique is exploited, in particular, to calculate the cardinality of Dezert and Smarandache. Finally, we applied this new methods to process the classification of a hight resolution images of a forest area.