基于稀疏和低秩约束的压缩感知图像重建算法.pdfVIP

第 31 卷第4 期 金陵科技学院学报 Vo 1. 31 , No .4 2015 年 12 月 JOURNAL OF JINLING INSTITUTE OF TECHNOLOGY Dec. ,2015 基于稀疏和低秩约束的压缩感知图像重建算法 朱俊1 ，陈长伟2 秘 (1.金陵科技学院计算机工程学院，江苏 南京 211169; 2. 南京晓庄学院信息工程学院，江苏 南京 211194) 摘 要z 压缩感知(Compressive sensing ,CS) 图像重建算法是 CS 图像获取问题的一个研究重点。针对基于低秩 先验的重建算法不能有效重建图像的局部几何特征，为了进一步提高图像的重建质量，在低秩先验的基础上，引 人稀疏约束作为图像额外的先验知识，建立了基于低秩和总变差约束的 CS 图像重建模型，并给出了相应的优化 算法。试验结果表明，与传统的稀疏性先验重建算法和最好的 NLR(Compressive sensing via nonlocal low-rank regl山rization) 算法相比，该试验所提的算法能够获得更好的图像重构效果。 关键词:压缩感知;图像重建;总变差;低秩属性 中图分类号:TP39 1. 41 文献标志码:A 文章编号: 1672 - 755X(2015)04 - 0023 - 04 A Compressive Sensing Image Reconstruction Algorithm ßased on Low-rank and Total Variation Regularization ZHU Jun 1 ,CHEN Chang-wei2 骨 (l.Jinling Institute of Technology ,Nanjing 211169 ,China; 2.Nanjing Xiaozhuang University ,Nanjing 211194 ,China) Abstract: Compressive sensing(CS)image reconstruction algorithm is a key point in the CS im- age acquisition. The low-rank reconstruction algorithms cannot eHectively reconstruct various image structures.ln order to improve the reconstruction precision with the same number of measurements , the sparsity regularization is introduced as the additional prior information of image , and a reconstruction model based on low-rank property and Total variation(TV)CS im- age has been proposed.We also propose the relative numerical algorithm in the paper.Compared with the traditional sparsity regularized algorithms and Compressive sensing via nonlocal low rank regularization(NLR) method which shows the state-of-the-art performance , the proposed algorithm can achieve better image rec