基于运动目标识别技术的视频监控系统  

Video Monitoring System Based on Moving Object Recognition

作　　者：高钰[1] 常永亮[2] 穆战松[1] 

机构地区：[1]安徽铜都铜业股份有限公司铜材厂,安徽铜陵244021 [2]中南大学信息科学与工程学院,湖南长沙410083

出　　处：《计算技术与自动化》2005年第4期29-31,共3页Computing Technology and Automation

摘　　要：设计了一种基于运动目标识别技术的数字视频监控系统,对生产现场有无工人作业进行判断,实现无人值守的智能化监控。系统采用客户机/服务器模型,利用企业局域网进行视频信号传输;利用帧间差分图像检测运动目标,实现视频运动对象的自动分割;采用边缘检测等技术提取运动目标的特征量;最后利用特征匹配算法对运动目标进行识别,判断目标类型,当现场长时间无工人作业时,发出报警信息。目前系统已投入运行,识别准确性较高。A digital video monitoring system based on moving object recognition is designed in this paper to judge whether there are workers in workshop and to realize intelligent monitoring system without Attenbant. The system adopts C/S model, utilize enterprise＇s LAaN to Transmit the video signals, utilize inter- frame gray difference to automatically separate the motion region from background, adopts edge detection technology to obtain the feature of moving object, and utilize the feature matching algorithm to recognize the moving object finally. While there are no workers in the workshop for a long time, the alert will be sent out. The system has been running on the spot, and the precision of recognition is relatively high.

关 键 词：视频监控系统 分割 特征匹配 目标识别 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]