基于DCS的WMSN多视角视频编解码  

Multi-view video coding of WMSN based on distributed compressed sensing

作　　者：罗晖[1] 祁美丽[1] 刘洁丽[1] 褚红亮[1] 王世昌[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：《计算机工程与设计》2013年第7期2492-2497,共6页Computer Engineering and Design

基　　金：国家自然科学基金项目(61261040/F010401)

摘　　要：为了有效解决无线多媒体传感器网络中多视角视频监控传输数据量大以及网络能量、资源受限的问题,提出了一种基于分布式压缩感知的高压缩率多视角视频编解码方法。对多视角视频序列进行分组处理,并将图像组分为关键帧和非关键帧;对关键帧采用基于压缩感知(compressed sensing,CS)的编解码方法进行处理;而在非关键帧的编码端采用联合稀疏表示方法对残差图像稀疏表示,解码端利用帧间时间相关性和多视角空间相关性预测生成当前视频帧,并借助差异补偿方法进一步提高预测准确性,同时提高了重构效果。实验结果表明,该方法取得较高的压缩率,重构出的图像质量比参考方法更高,且PSNR值得到了较大的提高。In order to solve the problems of transporting a large amount of multi-view video monitoring data and limited energy and resources in wireless multimedia sensor networks （WMSN） effectively, research a high compression rate of multi-view video coding method based on distributed compressive sensing theory. First, divide a group of video pictures into key frame and nonkey frames. Second, the key frame is coded based on CS method. In the processing of non-key frames, to sparse representation of residual images, the joint sparse representation method is used in the encoder. In the decoder, the current video frame is predicted using the temporal correlation between adjacent frames from a single view and the spatial correlation of multiple nearest views, and the accuracy of prediction is improved with disparity compensation methods, while the result of reconstruction is improved. The experiment results show that the method has higher compression ratio, and the reconstructed image quality is more efficiently than the reference algorithm, while PSNR are increased greatly.

关 键 词：分布式压缩感知 多视角 无线多媒体传感器网络 差异补偿 联合稀疏表示 

分 类 号：TN919.81[电子电信—通信与信息系统]