煤矿井下视频关键帧提取算法  

An algorithm of key frames extraction for video of coal mine underground

作　　者：程德强[1] 柳雪[1] 刘洋[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》2015年第7期54-57,共4页Journal Of Mine Automation

基　　金：国家自然科学基金项目(51204175);江苏省产学研前瞻性联合研究项目(BY2013019-6)

摘　　要：针对煤矿井下特殊的监控环境,对基于帧差欧氏距离的关键帧提取算法进行了研究。针对该算法存在的关键帧冗余度较大的问题,利用Canny算法提取图像边缘并进行边缘匹配,将冗余的关键帧剔除,从而降低冗余度。理论分析和仿真结果表明,改进后的关键帧提取算法能够有效降低冗余,提取出更加可靠的关键帧。In view of special monitoring environment of coal mine underground,algorithm of key frames extraction based on frame difference Euclidean distance was studied.For large redundancy of key frames of the algorithm,Canny algorithm was utilized to extract image edge,and edge matching was used to eliminate redundant key frames,so as to reduce redundancy.Theoretical analysis and simulation results show that the improved algorithm of key frames extraction can effectively reduce redundancy,and extract more reliable key frames.

关 键 词：关键帧提取 帧差 欧氏距离 CANNY算法 边缘匹配 

分 类 号：TD76[矿业工程—矿井通风与安全]