MULTI-FEATURE ADAPTIVE CLASSIFIERS FOR SAR IMAGE SEGMENTATION

We propose a multifeature scheme for terrain classification in SAR ima ge analysis, Different neural classifiers, trained on different featur es of the same sample space, are combined by using a non-linear ensemb le method. The feature extraction modules are chosen in order to disco ver the textural and contextual characteristics within the neighbourho od of each pixel. Comparisons with classical data fusion techniques an d consensus schema are reported.