Experimental comparison of neural networks for the classification of multisensor remote-sensing images

In this paper, we report the results of an investigation into the use of different neural models for the supervised classification of a multisensor (optical and radar) data set. We evaluated the performances of two well-known types of neural classifiers (i.e., MLPs, and Probabilistic Neural Networks (PNNs)) and compared them with the performances of the structured neural networks (SNNs) we proposed in [4, 5]. Further comparisons with the k-nearest neighbour classifier were also made in order to evaluate the validity of the considered neural networks as alternative classifiers to classical statistical ones.