A MULTIVARIATE GRADIENT AND MUTUAL INFORMATION MEASURE METHOD FOR HYPERSPECTRAL IMAGE VISUALIZATION

Hyperspectral imaging is becoming relevant in many applications. A large number of bands can be used to distinguish between different materials, quality inspection and mineral exploration. However, the information gathered by the narrow spectral bands in a hyperspectral image is not only large, but shows a high degree of correlation between bands. This redundancy needs to be reduced for efficient computation and storage. In this paper, we propose a new similarity metric, which combines gradient information and mutual information as positive interactive information. We apply the similarity metric for the visualization of hyperspectral images. Experimental methods show that the proposed method is superior to the state-of-the-art methods.