Comparison of the Multilayer Perceptron with Neuro-Fuzzy Techniques in the estimation of cover class mixture in remotely sensed data

Mixed pixels are a major source of inconvenience in the classifiaction of remotely sensed data. The paper compares MLP neural network with two neuo-fuzzy networks in the estimation of pixel component cover classes.The two neuro-fuzzy networks are: 1) the fuzzy multilayer perceptron (FMLP) and 2) a two stage hybrid (TSH) learning neural network, whose first stage consists of the FOSARt clustering network. Experimental results, obtained on a standard set of synthetic images, CLASSITEST, show that classification accuracies of FMLP and TSH are comparable, whereas TSH is faster to train then FMLP. In addition, FMLP and TSH outperform MLP when little prior knowledge is available for training the network.