A. Elmoataz et al. (Eds.): ICISP 2012, LNCS 7340, pp. 56–64, 2012. © Springer-Verlag Berlin Heidelberg 2012 Multisource Fusion/Classification Using ICM and DSmT with New Decision Rule Azeddine Elhassouny 1 , Soufiane Idbraim 1 , Aissam Bekkarri 1 , Driss Mammass 1 , and Danielle Ducrot 2 1 IRF-SIC Laboratory, Faculty of Science, Agadir, Morocoo {info_azeddine,soufianeidbraim,a_bekkari}@yahoo.fr, mammass@uiz.ac.ma 2 CESBIO Laboratory Toulouse, France Abstract. In this paper we introduce a new procedure for classification and change detection by the integration in a fusion process using hybrid DSmT model, both, the contextual information obtained from a supervised ICM classification with constraints and the temporal information with the use of two images taken at two different dates. Secondly, we have proposed a new decision rule based on the DSmP transformation, which is as an alternative and extension and overcoming the inherent limitations of the decision rules thus use the maximum of generalized belief functions. The approach is evaluated on two LANDSAT ETM+ images, the results are promising. Keywords: Detection of the changes, Image classification, Fusion, Hybrid DSmT model, Decision rule, DSmP, Satellite images, ICM. 1 Introduction The management and the follow-up of the rural areas evolution are one of the major concerns for country planning. The satellite images offer a rapid and economic access to accurate homogeneous and updated information of studied territories. An example of application which results from this is related to the topic of the changes cartography, in this paper, we are interested to study the most subtle changes of the Argan land cover and other themes in the region of Agadir (Morocco) by contextual fusion /classification multidates based on hybrid DSmT model [1-3] and ICM with constraints [4, 5]. Our work environment, is the theory of Dezert-Smarandache [1-3] which is recent and very little implemented or used before the covered work of this paper, it was applied in multidate fusion for the short-term prediction of the winter land cover [6-9] and, recently, for the fusion and the multidate classification [10-12], although the theory of evidence, it is more exploited for fusion/classification [12-18] also, for classifier fusion [19-20]. Our methodology can be summarized as following, after preprocessing of the images, a supervised ICM classification with constraints [4, 5] is applied to the two