Low-Complexity Multispectral Images Compression Algorithm Based Distributed Compressive Sensing

In this paper, we proposes a low-complexity and excellent multispectral images compression algorithm based distributed compressive sensing. 2-D lifting discrete wavelet transform (DWT) is applied to eliminate spatial redundancy of each band of multispectral images. Unlike the traditional wavelet-based coders (e.g., CCSDS-IDC, etc), DWT coefficients of each band here are not directly encoded, but the high-frequency sub-bands are re-sampled by a fast compressive sensing (CS) measurements. Then the resultant CS measurements of each band are encoded by means of distributed source coding. Experimental results show that the proposed compression algorithm obtains better compression performance compared with the relevant existing algorithms.