Hyperspectral Image Compression Based on Wavelets and Uniform Directional Filter Banks

Abstract

A multiscale geometric analysis method based on wavelet and uniform directional filter banks (WUDFB) is structured in order to overcome the insufficiencies of contourlet transform and wavelet-based contourlet transform which are used in compression. A novel WUDFB-based compression algorithm of hyperspectral imagery is proposed, in which intra-band WUDFB and inter-band prediction are used to remove the spatial and spectral redundancy respectively, and no-list set partitioned embedded block coder makes output embedded stream. The experimental results by using 64 band hyperspectral data show that the algorithm achieves higher peak signal-to-noise ratio at high compression ratio and preserves more detail information of original images than the algorithm based on wavelet transform.