Hyperspectral image analysis using neighborhood rough set and mathematical morphology

Classification of hyperspectral remote sensing images is a challenging task because of its high dimensionality. In this paper, a two stepped method for classification of hyperspectral images is presented. The first step performs band selection to reduce the `curse of dimensionality' issue by exploiting neighborhood rough set theory. The second step exploits mathematical morphology, in order to incorporate spatial information in the classification process. The result of the proposed method is compared with two state-of-the-art methods and found very much effective.