Lossless compression methods for hyperspectral images

Abstract

A novel efficient algorithm for lossless compression of hyperspectral images has been developed. The algorithm uses entropy function as a measure of interband similarity instead of correlation and mutual information functions that have been previously involved. The algorithm owes its performance to a newly invented adaptive prediction model. Image band reordering and prediction have been done according to the model steps. Edmond's algorithm has been proposed for finding optimal band ordering as an alternative to Prim's algorithm. The results obtained are outstanding. Moreover, three best lossless compression algorithms have been reviewed, tested using the same satellite data and compared to our method. We show that the proposed algorithm is capable of achieving compression ratios superior to that of the best-known lossless compression algorithms for hyperspectral images.