摘要
目前运动目标入侵检测的算法有很多种,一些算法已经解决了一些监控中的实际问题,但是还存在很多挑战性的难题没有解决,如针对小区视频智能监控系统的研究比较少。对小区视频智能监控系统中的入侵检测算法进行了研究,设计并实现了较稳定、实时性和扩展性较强的小区智能视频监控系统。通过引入减缓因子等从局部和全局光照变化两个方面改进了混合高斯背景模型的参数更新方案,提出了基于万有引力定律的边缘检测算法,并将其与对称差分相结合,有效地解决了运动目标检测结果存在空洞、边缘信息破碎的问题。
Currently there are many kinds of intrusion detection algorithms of the moving targets, some of which have solved some practical monitoring problems. But there are many challenging ones unsolved, and still fewer studies of community intelligent video surveillance system. Therefore, the intrusion detection algo- rithms of the neighborhood video intelligent monitoring system are examined, and the intelligent video moni- toring system with relative stability, real-time and strong scalability is designed and implemented. By introdu- cing a mitigating factor, and from two aspects like the local and global illumination change, the parameters- updating scheme of Gaussian mixture background model is improved. The edge detection algorithm based on the law of universal gravitation is put forward, which is combined with the symmetric difference, effectively solving the problems of the emptiness and brokenness of the edge information of the moving target detection results.
出处
《沈阳航空航天大学学报》
2015年第4期83-89,共7页
Journal of Shenyang Aerospace University
基金
沈阳航空航天大学青年教师自选科研课题(项目编号:201337Y)
关键词
智能视频监控
入侵检测
边缘检测
混合高斯背景
intelligent monitoring
intrusion detection
edge detection
gaussian mixture background
