人群运动方向异常检测算法

计算机科学 ›› 2013, Vol. 40 ›› Issue (Z11): 337-340.

人群运动方向异常检测算法

刘赏,董林芳

天津财经大学信息科学与技术系天津300222;天津财经大学信息科学与技术系天津300222

出版日期:2018-11-16 发布日期:2018-11-16
基金资助:
本文受天津市高等学校科技发展基金计划项目:基于视频分析的人群异常群体行为模式研究(20080816),天津市高等学校科技发展基金计划项目:基于无线视频的安全监护系统研究(20090809)资助

Abnormal Crowd Movement Direction Detection Algorithm

LIU Shang and DONG Lin-fang

Online:2018-11-16 Published:2018-11-16

摘要/Abstract

摘要： 运动方向是人群运动的一个重要特征。运动方向统一有序的人群运动中,人与人之间的受力小、存在碰撞的可能性低；而在方向杂乱的运动中,人与人之间受力较大、存在碰撞的可能性大,进而可能会导致踩踏等安全事故。因此,给出了一种新的人群运动方向异常检测方法,该算法利用光流法计算出人群的速度矩阵和运动方向矩阵,基于以上两个矩阵计算出“帧非同向运动指数”,并以此为依据来评价当前运动人群的运动是否存在异常。实验表明,“帧非同向运动指数” 直接体现了当前人群运动是否有序,因而基于运动方向的人群异常检测算法能够有效地检测出人群运动方向是否发生了混乱,以避免在方向杂乱的运动中发生危险事故。

关键词: 光流法,运动方向,人群运动异常,公共安全

Abstract: Direction is an important feature of crowd movement. The force between each other is small and the possibility of collision is low in orderly crowd movement,while they are higher in messy crowd movement which can leads security incidents,i.e. stampede. A new algorithm is proposed in this paper which detects the direction of abnormal crowd movement. First,this algorithm computes the velocity matrix and direction matrix by using optical follow method,and by these two matrices an index－“different movement index of frame” is computed.This index can indicate whether abnormal crowd phenomenon is appeared. The experiment results show that this index is relative with the messy degree of crowd movement. This algorithm can detect the abnormal movement effectively,which can avoid danger caused by messy movement.

Key words: Optical flow algorithm,Movement direction,Abnormal crowd movement,Public safety

刘赏,董林芳. 人群运动方向异常检测算法[J]. 计算机科学, 2013, 40(Z11): 337-340. https://doi.org/

LIU Shang and DONG Lin-fang. Abnormal Crowd Movement Direction Detection Algorithm[J]. Computer Science, 2013, 40(Z11): 337-340. https://doi.org/

参考文献

[1] Dirk H,Martin H.Derivation of a fundamental diagram for urban traffic flow[J]．European Physical Journal B,2008,70(2):229-241
[2] Wu H,Liang G,Lee K K,et al.Crowd Density Estimation Using Texture Analysis and Learning[C]∥IEEE International Conference on Robotics and Biomimetic．2006:214-219
[3] 柴斌.突发人群聚集事件智能视频监控[D].成都:电子科技大学,2011
[4] Bruhn A,Weickert J,Schnrr C.Lucas/Kanade Meets Horn/Schunck:Combining Local and Global Optic Flow Methods[J].International Journal of Computer Vision,2005,61(3):211-231
[5] Min Hu,Saad Ali,Mubarak Shah.Detecting Global Motion Patterns in Complex Videos[C]∥19th International Conference on Pattern Recognition (ICPR)．Tampa,2008,Florida,USA,1-5
[6] 徐珊.基于视频分析的异常群体事件检测[D].武汉:华中科技大学,2011
[7] 朱海龙,刘鹏,刘家锋,唐降龙.人群异常状态检测的图分析方法[J].自动化学报,2012,38(5):742-750
[8] Mehran R,Oyama A,Shah M.Abnomal corwd behavior detection using social force model[C]∥IEEE Conference on ComputerVision and Pattern Recognition,2009:935-942
[9] Ruan Jiu-Hong,Zhao Hua-an.Crowd movement segmentationusing velocity field histogram curve[C]∥Wavelet Analysis and Pattern Recognition 2012．2011:191-195
[10] Helbing D,Moln′ar P.Self-Organization Phenomena in Pedestrian Crowds[C]∥Self-Organization of Complex Structures:From Individual to Collective Dynamics.London Gordon and Breach．,1998:569-577
[11] 宁瑞芳,欧阳宁,莫建文.基于光流法的聚众事件检测[J].计算机工程与应用,2012,48(3):198-201
[12] Andrade E L,Blunsden S,Fisher R B． Hidden Markov Models for Optical Flow Analysis in Crowds[C]∥ICPR．2006(1):460-463

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed