Abstract: In order to implement the real-time surveillance and intelligent analysis in case of unattended, a novel intelligent surveillance system for campus was designed and developed via the software platform of VC6.0 and OpenCV1.0. The proposed system uses the GMM for the prospects detection, Meanshift algorithm for objects tracking, should track multi-target, and a new method based on the characteristic of HSV color space to avoid the shadow disturbance, then the system can be applied into some complex surveillance environments. The system can achieve the functions of intrusion detection, population statistics, crowd density detection and loiterer detection. With the experimental test in the actual campus, the results verify its usefulness and effectiveness. 

Key words: Video analysis, Intelligent surveillance, Security on campus, GMM, Shadow removal

摘要： 针对校园这一特殊环境，为了实现无人值守情况下的实时视频监控和智能分析，在ＶＣ６．０搭配ＯｐｅｎＣＶ１．０的软件平台上设计出一套基于视频分析技术的校园监控系统。本系统采用了ＧＭＭ 前景检测、基于Ｍｅａｎｓｈｉｆｔ算法的团块跟踪等多种实用性较强的视频分析技术，完成了对多目标的跟踪，并提出一种基于ＨＳＶ色彩空间特征的阴影消除新方法，使得系统可以应用于较复杂的监控环境。所设计的系统实现了入侵检测、人数统计、人员聚集检测和徘徊检测等功能，并在实际校园环境中进行了实际测试，结果证明了其实用性和有效性。

关键词: 视频分析 , 智能监控 , 校园安全 , 高斯混合背景建模 , 阴影去除