Lossless hyperspectral image compression using context-based conditional averages

In this paper, we propose a compression algorithm focused on the peculiarities of hyperspectral images. The spectral redundancy in hyperspectral images is exploited by using a context matching method driven by the correlation between adjacent bands of hyperspectral spectral images. The method compares favorably with recent proposed lossless compression algorithms in terms of compression, with significantly lower complexity.