大型追踪系统的多摄像头协同

Multi-camera coordination in large-scale intelligent tracking system

  • 摘要: 多摄像头间的目标追踪,即利用多摄像头协作完成对目标的监控追踪过程,是大规模智能监控系统中的重要研究方向。针对多摄像头追踪系统对多摄像头协同的实时高效性的需求,本文提出了一种利用多Agent代理多摄像头的分布式计算技术,单Agent层面分析其BDI模型,状态转移模型,多Agent层面分析其在目标追踪过程中完整的通信过程,以及多个摄像头的最佳状态选择。通过开源JADE框架,对基于多Agent的多摄像头追踪系统进行系统搭建,系统实现了多摄像头协同下对目标的持续追踪,实验证实了本文的多Agent系统可有效的实现多摄像头的协同,将通信延迟时间缩小到最短,以便及时更新每个摄像头的状态,从而大大提高了目标追踪的准确率,以及计算资源的利用率,达到对目标进行持续追踪的目的。

     

    Abstract: Object tracking among multiple cameras, that is, using multi-camera cooperation to achieve the goal of monitoring and tracking process, is an important research direction of large-scale intelligent surveillance system. The tracking performance strongly relies on the great results of multi-camera coordination. In this paper, a distributed framework with multi-agent instead of multi-camera was proposed based on the demand of multi-camera coordination, which is for research of BDI and state transfer model of single agent, the entire communication process of multi-agent tracking and the method how multi-camera control their state were analyzed. Multi-agent system was established by use of JADE. Experiments prove that the multi-agent system in this paper can realize the multi-camera coordination effectively, the communication delay time has been reduced to the minimum, in order to update the status of each camera rapidly, which greatly improves the accuracy of target tracking, and the utilization of computing resources, and the purpose of target tracking continuously has been realized.

     

/

返回文章
返回