Using the significant spectral correlation within the hyperspectral images
a lossless compression algorithm is presented in this paper.The method can be explained in two steps:(1)band ordering according to spectral correlation coefficient;(2)3-D adaptive prediction based on neural networks.Experiments show that this method can remove the spatial and spectral redundancy efficiently.The proposed approach can reduce about 0.3 bit per pixel compression compared with JPEG-LS.