Hyperspectral image classification using Daisy descriptor

In this paper it is proposed to use the Daisy descriptor approach to increase classification accuracy in Hyperspectral images. In this descriptor approach, a new descriptor is constructed by applying the spectral data to multiple low-pass filters with different pass frequencies. This approach was applied in the spatial domain to visible range images before; and is in this paper applied in the spectral domain to hyperspectral images for the first time. It is shown in this paper that the classification accuracy is increased in hyperspectral images, using nearest neighbor classification.