基于树型结构的摄像头网络智能监控系统设计

摘要

Video surveillance is an important part of public security. As the scale of surveillance network keeps growing, traditional surveillance systems fail to perform intelligent task on camera network. Large scale camera network requires high processing power and high efficient algorithm of correlation analysis. In order to solve this problem, a novel architecture for surveillance system is proposed, based on tree structure. This system can take full advantage of server cluster to efficiently run various algorithms and easily handle multi-camera correlation analysis loaded on it. This system has been tested in the network consisting of 20 HD cameras, and its stable performance verifies the rationalization of the design.%视频监控是安防的重要组成部分,随着摄像头监控网规模的增大,传统的监控系统已经无法胜任基于网络的智能监控任务.大规模摄像头网络要求高性能的处理能力和高效的联合分析智能算法,为了解决这个问题,提出了一种全新的监控系统架构:系统基于树型结构.它充分利用集群计算机的处理能力,高效地运行加载于其上的多种智能算法,便捷地进行多摄像头信息联合分析.系统已在20台高清摄像头组成的网络下实际运行,其稳定的性能证明了设计的合理性.