Spectral and spatial classification of hyperspectral data using SVMs and Gabor textures

High spectral and spatial resolution images are widely used in urban mapping. In order to increase conventional spectral classification accuracy, morphological profiles based spatial information were widely studied. Some improvements were obtained in terms of classification accuracy. However, texture is is a very useful structure information in high spatial resolution images. In this paper, Gabor filters based textures are investigated to increase conventional spectral classification. Gabor textures are extracted from the first three PCs of the hyperspectral bands. Then Gabor textures and hyperspectral bands are concatenated, and classified in one support vector machines. The results show that Gabor textures are useful in extracting spatial information and increasing the classification accuracy compared to conventional spectral classification.