公共场所人群加速度异常检测系统

发布时间：2018-10-05 15:12

【摘要】：针对目前基于速度检测公共场所密集人群异常行为存在的检测准确率低、使用范围局限的问题,从人群的加速度角度对可能导致公共安全事故的人群异常行为进行研究,提出了一种基于加速度检测人群异常行为的算法,并基于该算法实现了针对人群逃散、人群聚集、人群拥挤和人群逆行4种异常行为检测的系统。首先,利用金字塔Lucas-Kanade光流法进行特征点跟踪;然后,在获取到特征点的速度矩阵基础上进一步计算其加速度矩阵,反映速度的整体变化;最后,从加速度大小和方向两方面检测人群异常行为。结果表明,所提算法检测用时较少,相比基于速度检测的对比算法,检测的正确率提高到80%,误报率降低为5%。
[Abstract]:In view of the problem of low detection accuracy and limited range of use in detecting abnormal behavior of dense crowd in public places based on speed measurement, this paper studies the abnormal behavior of people who may cause public safety accidents from the point of view of crowd acceleration. An algorithm based on acceleration is proposed to detect abnormal behavior of crowd. Based on this algorithm, four detection systems of abnormal behavior such as escaping, crowd gathering, crowd crowding and crowd retrograde are implemented. Firstly, the pyramid Lucas-Kanade optical flow method is used to track the feature points. Then, the acceleration matrix is further calculated on the basis of obtaining the velocity matrix of the feature points to reflect the global variation of the velocity. The abnormal behavior of the crowd was detected in terms of acceleration magnitude and direction. The results show that the detection time of the proposed algorithm is less, compared with the contrast algorithm based on speed detection, the accuracy of the detection is increased to 80 and the false alarm rate is reduced to 5 percent.
【作者单位】：天津财经大学理工学院;
【基金】：国家自然科学基金项目(61502331) 天津市应用基础与前沿技术研究计划项目(15JCQNJC00800) 中国民航信息技术科研基地开放课题(CAAC-ITRB-201504)
【分类号】：C913

【相似文献】