Camouflage assessment based on hyperspectral characteristics

Camouflage assessment is an essential part in camouflage design and identification. In order to make an effective camouflage evaluation under the hyperspectral detection, we presented a new method based on hyperspectral characteristics considering the characteristics of high spectral resolution of hyperspectral images. Firstly, the spectral information was extracted and analyzed in spectral dimension. Then, the multi-wavelet decomposition was used in spatial dimension of the first principal components to extract texture information such as the ratio of high frequency component and low frequency component, ratio of horizontal component and vertical component, contrast, correlation, energy and homogeneity of low frequency image. Finally, Euclidean distance was calculated to reflect the similarity of spectrum and texture. Experimental data demonstrates the objectivity of the approach. The comprehensive distance between camouflage raincoat and lawn background is 1.3225 that proves they fuse best.