Unsupervised hyperspectral image segmentation using adaptive bilateral filtering

This paper proposes the use of adaptive bilateral filter for the segmentation of hyperspectral images. First, the spectral bands are selected according to the information contained in each band. Then on each band, adaptive bilateral filter is applied in order to increase the spatial correlation of each pixel. The results are evaluated based on the successful segmentation percentage. It is shown that the segmentation accuracy of k-means clustering algorithm is increased.