基于奇异值分解的数字图像压缩技术研究

摘要

In order to realize the image compression, on the basis of analyzing image compression principle, a matrix singular value decomposition(SVD) image compression algorithm has been proposed. Based ondigital image matrix singular value decomposition was made in the algorithm, an image was converted into singular value matrix containing several nonzero value, and the image compression can be realized.Bymatlab simulation experiment, the singular value varying from 0 to 240, When the singularvalues is greaterthan50,with the increase of singular value the compression ratio is smaller and smaller, and the image become more and more clear.Compared with the original image, using the singular value decomposition of matrix compression method, the original image can be compressed by about 20%, which has good compression performance.%为了实现图像压缩,在分析图像压缩原理的基础上,提出了一种矩阵奇异值分解(SVD)的图像压缩算法,该算法通过对数字图像矩阵进行奇异值分解,将一幅图像转换成包含几个非零值的奇异值矩阵,从而实现了图像压缩.通过Matlab仿真实验,在奇异值从0变化到240的过程中,当奇异值大于50时,随着奇异值的增大,压缩比越来越小,图像慢慢变清晰.和原始图像相比,采用矩阵的奇异值分解压缩方法可以将原始图像压缩20%左右,具有较好的压缩性能.