基于模糊推理的视频监控方法  

Video Monitoring Method Based on Fuzzy Reasoning

作　　者：赵栋栋[1] 冯永浩[1] 张斌[1] 

机构地区：[1]空军工程大学信息与导航学院,西安710077

出　　处：《科学技术与工程》2013年第28期8315-8319,共5页Science Technology and Engineering

摘　　要：针对拥挤情形下的人员目标跟踪问题,提出了一种使用模糊推理机制的视频监控新方法。首先采用背景减除法获得目标的前景图像,再利用特征搜索法去测量目标的新的位置变化,然后使用混合的模糊信息推理机制的方法去验证跟踪结果,并且根据需要对误差进行补偿。针对不同场景进行了实验,实验显示,对误差进行补偿之后,在拥挤情形下的召回率和准确率都能达到90%以上,平均误差降低了10%以上。实验结果证明,能够有效提高视频监控系统对于拥挤情形下多目标跟踪的性能。Aim at the problem of congestion situations staff target tracking, a new video surveillance method rely on fuzzy inference mechanism is put forward. Firstly, background subtraction method is used foreground image, and use d characteristics search method to measure new position change to obtain the objectiveg of target, according to the need to compensate the error. Experiments were carried out for different scenarios. Experiments show, use this method to compensate the error, the recall rate and precision rate can reach more than 90% in crowded situations, the average error is reduced by more than 10%. The experimental results show that this method can effectively im- prove the performance of video surveillance system for multiple target tracking in congestion situations.

关 键 词：视频监控 模糊推理 误差补偿 人员遮挡 目标跟踪 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]