一种基于粒子轨迹的密集人群跟踪方法

Visual surveillance based crowd behavior analysis is currently a highly focused research area of intelligent visual surveillance technology. The primary problem of crowd behavior analysis is target detection and crowd tracking. In this paper, we present a method for crowd tracking in which crowds are treated as collections of particles by analysing the characteristic of the moving crowd. In the algorithm, the flow map is calculated from the video by Lucas-Kanade optical flow algorithm, then the crowd tracking is completed based on the optical flow and particle tracking algorithm. Experimental results verify the effectiveness of this approach on several open datasets.