基于运动检测的变电站遥视技术  

Motion Detection Based Remote Video Monitoring Technique in Substations

作　　者：郑连清[1] 曾朱文[1] 

机构地区：[1]重庆大学电气工程学院,重庆市400044

出　　处：《电力系统自动化》2005年第22期77-80,共4页Automation of Electric Power Systems

摘　　要：针对当前无人值守变电站遥视系统由人工监视的弊端,提出了在无人值守变电站遥视系统中加入视频分析的新构想,以提高监控水平和效果。在比较了当前多种常用视频分析算法优劣的基础上,择优选用了一种基于柯西分布统计模型的运动检测算法,并对模型的建立和维护进行了改进;采用ARM+Linux构成嵌入式系统进行了实验,实验结果表明这种系统具有较好的效果。Aiming at the defects in current remote supervision of unmanned substations, a new idea by introducing visual analysis into remote supervision of unmanned substations to improve the level of surveillance is presented. By comparing the advantages and disadvantages of some commonly used methods for visual analysis, a motion detection algorithm based on the Cauchy statistical model is selected and the definition and maintenance of the coefficients of the model are improved. The results of experiment using an embedded system constructed by ARM＋ Linux show that the system has fairly good performance.

关 键 词：无人值守变电站 遥视系统 运动检测 背景模型 参数估计 

分 类 号：TM764[电气工程—电力系统及自动化]