Remote Sensing Image Compression Based on the Improvement of SPECK Algorithm

Abstract

We conduct MS-TAWCP algorithm to promote the effectiveness of encoding. Although the SPECK algorithm has introduced an embedded block method, the algorithm itself ignores the relationship between wavelet coefficients in every level of coefficients, such as the characteristics of the coefficient correlation. The MS-TAWCP algorithm makes full use of the characteristics of coefficient correlation and employs double threshold to optimize the distribution of coefficient. Moreover, MS-TAWCP algorithm splits the wavelet coefficients into three frequency space, where employ different threshold for different frequency space, to avoid coefficient double search and duplication of judgment in the same threshold plane caused by double threshold. The experiment shows MS-TAWCP algorithm outperforms the SPECK about 3 db in PSNR, without comprising remote sensing image information and reducing algorithm speed.