基于边信息的分布式视频压缩感知重建.pdfVIP

• 58 • 《测控技术~2016 年第 35 卷第 2 期 基于边信息的分布式视频压缩感知重建 袁静1 ， 2 (1.宿迁学院计算机系，江苏宿迁 2238∞; 2. 南京邮电大学信号处理与传输研究院，江苏南京 21创)()3 ) 摘要:WZ 视频编码系统中的虚拟信道会因信道模型参数估计的不稳定导致系统性能下降，为了克服该 问题，提出不带有虚拟信道的分布式视频压缩感知重建算法，仅靠解码端获得的压缩感知观测值对非关 键帧的错误进行纠正，并利用平滑投影Landweber 迭代求解联合重建模型。实验证明本文所提出算法 获得了更好的率失真性能和更低的重建时间。 关键词:压缩感知;平滑投影;分布式视，频编码;边信息;联合重建 中固分类号:TN91 1. 73 文献标识码:A 文章编号:l(削 -8829(2016)02 -0058-04 Distributed Compressed Video Sensing Reconstruction Based on Edge Information l YUAN Jing , 2 ( 1. Computer Department, Suqian College , Suqian 2238∞， China; 2.Institute of Signal Processing and Transmission, Nanjing University of Posts and Telecommunications, Nanjing 21α则， China) Abstract: TO sOlve the problem Of inaccurate estimatiOn tO由e parameters Of virtual channel in the Wyner-Ziv (WZ) videO cOdec system, the smoothed projected distributed cOmpressed videO sensing recOnstructiOn algO- rithm is propOsed. 咀le method remOves the virtual channel and direcÙY u肥s cOmpressed sensing (CS) me部­ urements tO cOrrect eπors Of nOn key frame , and uses sm∞也 projectiOn Landweber iterative tO sOlve j Oint re- cOnstructiOn model. Experiments shOw 由at the propO sed algOrithm has ever better rate-distOrtiOn perfOrmance and lOwer recOnstructiOn time. Key words: cOmpressed sensing; smOOthed projection; distributed videO cOding; edge infOrmatiOn; j Oint recOn- struchOn 分布式视频压缩感知系统基本借鉴了 Benrd Gi- cOmpressed sensing) 理论，其分析并证明了压缩感知与 rod 等人提出的 WZ 视频编码系统，将视频流划分为非 分布式编码融合