无线传感器网络中基于压缩感知的静止图像压缩方案研究  

The Research on Compressive Sensing Based Still Image Compression in Wireless Sensor Network

作　　者：李然[1] 武明虎[1] 郑海波[1] 崔子冠[1] 朱秀昌[1] 

机构地区：[1]南京邮电大学江苏省图像处理与图像通信重点实验室,江苏南京210003

出　　处：《南京邮电大学学报（自然科学版）》2013年第4期44-49,54,共7页Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition

基　　金：国家自然科学基金(61071091,61271240);江苏省研究生科研创新计划(CXZZ12_0466,CXZZ11_0390);江苏省高校自然科学研究(12KJB510019);南京邮电大学校科研基金(NY212015);湖北省教育厅科研重点(D20121408)资助项目

摘　　要：受功耗、成本和无线传输环境的限制,无线传感器网络中采集图像的节点并不适合采用传统静止图像压缩标准JPEG、JPEG2000。为了适应无线传感器网络的"轻编码、重解码"的特点,提出了一种基于压缩感知(Compressive Sensing,CS)的静止图像压缩方案。该方案在编码端进行分块测量,分块测量次数根据图像的边缘信息自适应地分配;在解码端则利用分块观测数据进行整帧重建。仿真实验表明,与已有的基于CS的图像压缩方案相比,该方案可有效地降低CS测量成本,并可在重建时消除图像的边缘模糊与块效应现象。The traditional still image compression standard JPEG and JPEG2000 are not appropriate for a camera sensor in the wireless sensor network since it is limited by the power consumption, costs and wire- less transmission environment. In order to meet the ＂light encoding and heavy decoding＂ characteristic of the wireless sensor network, this paper proposes a novel still image compression scheme based on the compressive sensing （CS） . The scheme measures each image block at the eneoder and adaptively assign the number of measurement for each block in terms of their edge information. At the decoder, the scheme uses the measurements of each block to reconstruct the whole image. Simulation results show that the pro- posed scheme can effectively reduce the cost of CS measuring in comparison with the existing CS based image compression schemes, and it can also remove edge blurs and blocking artifacts in the reconstructed image.

关 键 词：压缩感知 无线传感器网络 静止图像压缩 图像重建 分块自适应测量 平滑投影Landweber算法 

分 类 号：TN919.8[电子电信—通信与信息系统] TN911.73[电子电信—信息与通信工程]