基于块分类的矿井视频图像DCVS重构算法  

Reconstruction algorithm DCVS for mine video image based on block classification

作　　者：赵小虎 沈雪茹[1] 李婉梅 王宽 ZHAO Xiaohu;SHEN Xueru;LI Wanmei;WANG Kuan(School of Information & Control Engineering,China University of Mining & Technology,Xuzhou 221008, China;Research Center of Internet of Things ( Perception Mine),China University of Mining & Technology,Xuzhou 221008, China;National & Local Joint Engineering Laboratory of Mine Internet Application Technology,China University of Mining & Technology,Xuzhou 221008, China)

机构地区：[1]中国矿业大学信息与控制工程学院,江苏徐州221008 [2]中国矿业大学物联网(感知矿山)研究中心,江苏徐州221008 [3]中国矿业大学矿山互联网应用技术国家地方联合工程实验室,江苏徐州221008

出　　处：《扬州大学学报（自然科学版）》2019年第2期70-74,共5页Journal of Yangzhou University：Natural Science Edition

基　　金：国家重点研发计划资助项目(2017YFC0804404)

摘　　要：针对煤矿井下视频监控图像因数据量大而导致传输和存储困难等问题,引入压缩感知理论对视频图像进行编解码,提出一种新的分布式视频编解码(distributed compressed video sensing, DCVS)方案.为了获得更稀疏的表示和更普遍的适用性,提出一种基于块的自适应混合稀疏基方案.针对边信息获取过程中通常采用固定权值合成边信息而忽略不同图像块之间相关性的问题,提出一种块分类加权边信息生成方案.结果表明:与传统的视频编解码方案相比,基于块的分类编解码方案能充分利用帧间相关性,在不同采样率下视频重构的峰值信噪比均有所提高,视频帧重构的质量也得到有效提升.In view of the problem that the video monitoring image of coal mine underground is difficult to transmit and store due to the large amount of data, the theory of compressed sensing is introduced to codec video images, and a new distributed video coding and decoding scheme is proposed. In order to obtain a more sparse representation and more general applicability, a block-based adaptive hybrid sparse basis scheme is proposed. Aiming at the problem that the side information is usually combined with fixed weights and the correlation between different image blocks is neglected, a block classification weighted side information generation scheme is proposed. The results show that compared with the traditional video codec scheme, the block-based classification and decoding scheme can make full use of the inter-frame correlation. At different sampling rates, the peak signal-to-noise ratio of video reconstruction is improved, and the quality of video frame reconstruction is effectively improved.

关 键 词：块分类 压缩感知 分布式视频编解码 稀疏表示 边信息重构 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]