Wavelet networks for hyperspectral images classification

The idea of using artificial neural network has proved useful for hyperspectral image classification. However, the high dimensionality of hyperspectral images usually leads to the failure of constructing an effective neural network classifier. To improve the performance of neural network classifier, wavelet-based feature extraction algorithms are applied to extract useful features for hyperspectral image classification. Besides, the wavelet networks which integrates the advantages of wavelet-based feature extraction and neural networks classification has been proposed with some success in data classification. In this paper, the neural networks, wavelet networks and wavelet-based feature extraction were firstly described and then a hyperspectral data set from AVIRIS was used to test the effectiveness and performance of classification using these three methods. The experiment results showed that the wavelet networks has better classification accuracy than traditional back propagation neural networks, and exactly is an effective tool for classification of hyperspectral images.