交通监控中基于压缩感知的WMSN视频编解码  

WMSN Video Encoding/Decoding Based on Compressed Sensing in Traffic Surveillance

作　　者：罗晖[1] 王世昌[1] 褚红亮[1] 杨成武[1] 

机构地区：[1]华东交通大学信息工程学院,江西南昌330013

出　　处：《华东交通大学学报》2013年第1期52-57,共6页Journal of East China Jiaotong University

基　　金：国家自然科学基金项目(61261040)

摘　　要：在交通监控中,利用WMSN节点获取路况视频,适用于复杂多变的地理环境,但视频图像数据量大,而WMSN节点资源受限。基于压缩感知技术和传统视频编码理论,提出了一种新的WMSN视频编解码方法。在编码端,依据视频图像的相关性对视频流进行分帧,再根据其特点分别进行压缩感知处理。在解码端,运用OMP算法重构关键帧和残差帧视频图像,进而恢复视频流。仿真实验表明,在保证视频图像重构质量的前提下,可以减少监控网中视频传输的数据量。In traffic surveillance,WMSN is applied to capture the video of road state,which can adjust to complex and variable environments.But WMSN sensor’s resource is limited and the number of video images yielded is tremendous yet.Based on compressed sensing and traditional video coding theory,this paper proposes a new method of WMSN video encoding/decoding.In encoding,the video stream is divided into key frames and non-key frames according to video image correlation,which are processed in line with their characteristics respectively.In decoding,video stream is further reconstructed after the key frames and residual video images are reconstructed through OMP algorithm.The simulation experiments demonstrate that the data volume of video transmission is largely reduced on the premise of keeping the reconstructed video image quality.

关 键 词：交通监控 WMSN 压缩感知 视频编解码 

分 类 号：TN919.8[电子电信—通信与信息系统]