Data fusion and feature extraction in the wavelet domain

This paper concentrates on data fusion, feature extraction, feature selection and neural network classification for multi-source remote sensing and geographical data. The considered feature extraction method is based on the discrete wavelet transformation (DWT). The original data are transformed using DWT and then a feature selection mechanism is applied to select features from the full feature set in the wavelet domain. The feature selection mechanism is a binary genetic algorithm which selects the best features to be used in a neural network classification. In experiments on two datasets, the proposed data fusion and feature extraction method performed well in terms of overall accuracies as compared to results obtained with other well-known feature extraction methods.