A comparative study of radial basis function neural networks and wavelet neural networks in classification of remotely sensed data

Artificial neural networks (ANN) constitute a powerful class of nonlinear function approximate for model-free estimation. ANN has been widely used in pattern recognition, prediction and classification. In the artificial neural network approach, we compare radial basis function neural networks (RBFNN) and wavelet neural networks for multispectral image classification. The aim of this study is to examine the effectiveness of the neural network model for multispectral image classification. Radial basis function neural network is used for its advantages of rapid training, generality and simplicity over feedforward backpropagation neural network. The k-means clustering is used to choose the initial radial basis centers and widths for the RBFNN. The wavelet is a localized function that is capable of detecting some features in signals. A wavelet basis function is assigned for each neuron and each synaptic weight is determined by learning. An attempt is also made to study the performance of the RBFNN with the centers and widths chosen using the classical k-means clustering.