Comparative study of information fusion methods for sonar images classification

Here, a comparative study of information fusion methods for sonar images classification is proposed. The automatic classification of sonar images is a very difficult problem. Our first task consists in finding a good image representation to classify the sea bottom. Classical approaches are based on texture analysis. Many methods can be considered to deal with this problem; however, the best choice of the considered method depends often on the kind of sediment. Once the features extraction method has been considered, many classifiers can be used. In order to extract features, four major texture analysis methods have been considered. The four sets of features are classified and different methods of information fusion, such as the weighted vote approach, or coming from the possibility theory and evidence theory, have been employed.