一种基于小波压缩感知的藏族壁画图像处理  

A Tibetan Mural Image Processing Based on Wavelet Compressed Sensing

作　　者：孙燕[1] 李晓光[2] 卓力[2] 潘春花[1] 朱存[1] 

机构地区：[1]青海民族大学计算机学院,青海西宁810007 [2]北京工业大学信号与信息处理实验室,北京100124

出　　处：《实验室研究与探索》2016年第5期138-140,216,共4页Research and Exploration In Laboratory

基　　金：青海民族大学与北京工业大学基础研究合作项目

摘　　要：针对藏族壁画特点在采集图像时采用一种基于小波压缩的感知算法,利用改进的小波变换,分解图像高频分量和低频分量,并进行稀疏变换,达到采样压缩。图像重构利用压缩感知OMP算法重构图像,实验表明,原始图像分块数和随机矩阵不变时,采样率越大,PSNR越小,重构的图像也越清晰,针对藏族壁画算法有效。According to the feature of Tibetan murals,a compressed sensing algorithm is presented based on wavelet.The improved compressed sampling algorithm separates the high frequency sub-bands from the low frequency sub-band,and realizes sparse transformation. The compressed sensing OMP algorithm is employed in image reconstruction.Experiments show if the number of blocks and random matrix is fixed,then the sampling rate is higher,the PSNR will be smaller,and the reconstructed image will have better quality. Hence the algorithm is effective for Tibetan murals.

关 键 词：藏族壁画 小波压缩感知 OMP 算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]