论文部分内容阅读
针对当前公共安全监控系统在应对突发事件时,因监视器数量多、信息量大、事件变化快等因素而导致的不能及时有效地对移动对象进行跟踪的问题,提出了一种基于拓扑结构的多摄像头移动对象辅助协同跟踪方法。首先将城市路网中的摄像头监控区域抽象为虚拟节点定位至结点弧段模型中,构建摄像头网络拓扑结构;根据拓扑关系搜索出与发现移动对象的摄像头关联的摄像头集合;再分析摄像头间的空间邻近关系和时间差异,获取时序关系最优的摄像头组作为重点监控区域进行移动对象跟踪,并定期进行协同更新。最后建立了一个模拟监控系统验证了利用多摄像头协同方法实现对移动对象实时跟踪的有效性。
Aiming at the problem that the current public safety monitoring system can not timely and effectively track the moving objects due to many factors such as the number of monitors, the large amount of information and the rapid change of events when dealing with unexpected events, a topology- Multi-camera mobile object-assisted collaborative tracking method. Firstly, the surveillance area of the camera in the urban road network is abstracted as the location of the virtual node in the arc segment model of the node, and the topological structure of the camera is constructed. According to the topological relationship, the camera set associated with the camera that discovers the moving object is searched out. Spatial proximity and time difference, the camera group with the best timing relationship is taken as the tracking area for tracking moving objects, and the collaborative updating is performed periodically. Finally, an analog monitoring system was established to validate the effectiveness of real-time tracking of moving objects by using multi-camera coordination method.