摘要
伴随视频监控系统的普及,多摄像机智能视频监控技术得到了广泛的研究和应用。特别是视频监控中的运动目标检测分割、运动目标跟踪和多摄像机协同技术,成为计算机视觉领域研究的热点。单摄像机跟踪常被应用的监控范围限制,当监控场景是一个室外环境或需要一个较大的观测范围时,常需要多摄像机进行多物体跟踪。为了建立一套有效的多摄像机智能监控系统,提出一种基于摄像机视野的重叠区域的协同监控方法,找出重叠区域的边界线即场边界线,旨在解决单摄像机监控范围有限的问题,实现对某一场景进行全景监控。在多摄像机系统中各个摄像机的场边界线在其他摄像机中可见的情况下,给出寻找它们的方法。基于此,可基本确定在多摄像机中的运动目标对应关系,获取多摄像机的拓扑关系。
With the popularization of video monitoring system,the multi-camera intelligent video monitoring technology has been widely studied and applied. The moving object detection and segmentation,moving target tracking,and multi-camera cooperation technology have become the research focus in computer vision field. Single camera tracking is limited by the monitoring scope. When the monitoring scene is an outdoor environment or needing a larger monitoring scope,multi-camera are required to track multi-object. A coordination monitoring method of overlapping region based on camera visual field is proposed to establish an effective multi-camera intelligent monitoring system. The field of view（FOV）lines are found out in the overlapping region to solve the limited monitoring scope of the single camera. The panorama monitoring for a certain scene is realized. The way how to find the FOV lines in multi-camera system is provided when each camera＇s FOV lines can be seen by other cameras. The corresponding relationships of moving object in multi-camera can be ensured basically,and the topological relation of multi-camera can be acquired.
出处
《现代电子技术》
北大核心
2015年第21期6-10,共5页
Modern Electronics Technique
基金
江苏省自然科学基金(BK20131280)
南京大学金陵学院教学改革资金项目(0010111208)
关键词
运动目标检测
多摄像机协同
场边界线
全景监控
moving object detection
multi-camera cooperation
field of view line
panorama monitoring
