基于图像分块的Toeplitz结构测量矩阵设计  

Design of Toeplitz Structure Measurement Matrix Based on Image Blocking

作　　者：瞿广财[1] 张淑芳[1] 吕卫[1] 褚晶辉[1] 

机构地区：[1]天津大学电子信息工程学院,天津300072

出　　处：《计算机工程》2012年第16期212-214,218,共4页Computer Engineering

基　　金：国家自然科学基金资助项目(61101226);内燃机燃烧学国家重点实验室开放课题基金资助项目

摘　　要：现有压缩感知成像系统存储测量矩阵时需要较大空间,针对该问题,提出一种基于图像分块的Toeplitz结构块循环测量矩阵设计方法。将图像分块进行压缩感知,减少测量系统的存储空间,从而降低硬件实现难度。仿真结果表明,该方法能快速有效地获得测量值,且重构图像的主客观质量较好。According to the problem of existing Compressed Sensing（CS） imaging system requiring a large storage space to store the measure- ment matrix, a Toeplitz structure block-circulant measurements matrix design method based on image blocking is proposed in this paper. It uses the image block compressed sensing theory, decreases the storage space of the measurement system efficiently and reduces the difficulty of hardware implementation. Simulation results show that the proposed method can acquire the measurements quickly and effectively, and the image reconstructed has little loss in subjective and objective quality.

关 键 词：压缩感知 测量矩阵 Toeplitz结构 图像分块 块循环矩阵 

分 类 号：TN911.73[电子电信—通信与信息系统]