ABSTRACT
This chapter describes a preprocessing step needed to make use of incomplete Geographical Information Systems databases as a priori information. It describes the data fusion system designed for land-use mapping and activity monitoring using evidential fusion. The chapter shows the results obtained using the data set provided by the Canadian Space Agency as part of the second installment of the Application Development and Research Opportunity program. It provides some of the first results obtained with the proposed system prototype. Its capability to perform change detection/identification has been evaluated on a set of seven RADARSAT-1 scenes acquired over the city of Stephenville and its vicinity. In genetic algorithms, the fitness value is a measure of the performance of a chromosome. The chapter aims to improve the texture classification rate. The classification map is computed using the features selected by each chromosome.
