摘要
为了解决变电站视频监控系统中大量视频图像的困扰,在系统中采用了行为识别技术。将输入的压缩视频流解码,生成图像数据,再对图像数据进行识别。识别内容包括:指定区域是否存在运动物体;是否有物体越过警戒线;进入变电站的工作人员是否戴红色安全帽。如果当前图像存在上述某一个或多个行为的检测信息,则在监视界面上给出提示,并采用多种方式处理相应信息。实际应用表明,系统实现效果达到了设计的期望。
In order to solve the puzzle of large video data from video monitoring system, the behavior identifying technique is adopted in the system. The input compressed video stream is decoded to image data, which is to be identified then. The identifying items include： ①Is there any moving object existing in specific area？ ②Does any moving object run across the alert line？ ③Whether worker enters into substation with red helmet or not？ If one or several identifying behaviors happen on image, alert information will be shown on the monitoring interface, and relevant information will be processed by many ways. Practice shows that the system has satisfied design expectation.
出处
《电力系统自动化》
EI
CSCD
北大核心
2010年第22期117-119,共3页
Automation of Electric Power Systems
基金
福建省电力有限公司科技项目资助(NC200950)
关键词
变电站
视频监控系统
行为识别
substation
video monitoring system
behavior identification
