基于全搜索的煤矿井下视频序列边信息生成算法  

An algorithm of side information generation of coal mine underground video sequence based on full search

作　　者：程德强[1] 刘洁[1] 郭政[1] 

机构地区：[1]中国矿业大学信息与电气工程学院,江苏徐州221008

出　　处：《工矿自动化》2014年第10期19-22,共4页Journal Of Mine Automation

基　　金：国家自然科学基金项目(51204175)

摘　　要：针对煤矿井下视频分布式编码中图像照度低和清晰度低等特点,提出了基于全搜索的煤矿井下视频序列边信息生成算法。该算法对已解码的关键帧进行扩编,并对全搜索时每个参考帧对应像素块的均方误差函数值通过阈值进行分析和判断,保留低于阈值的运动矢量,提高边信息生成质量。理论分析和仿真结果表明,本文提出的算法可以实现煤矿井下视频序列的高质量边信息的生成,为将分布式编码广泛应用于煤矿井下视频监控和通信提供了基础。In view of characteristics of low illumination and low definition of image in distributed coding of coal mine underground video,an algorithm of side information generation of coal mine underground video sequence based on full search was proposed.The algorithm expanded key frames which have been decoded,analyzed and judged the mean square error function value of corresponding pixel of block of each reference frame during full search by threshold.It reserved motion vector below the threshold,and improved quality of side information generation.The theoretical analysis and simulation results show that the algorithm can realize generation of high-quality side information of coal mine underground video sequence,and provides foundation for distributed coding widely used in coal mine underground video surveillance and communication.

关 键 词：煤矿井下视频 全搜索 视频序列 边信息 

分 类 号：TD673[矿业工程—矿山机电]