APP下载

基于选定区域颜色直方图的粒子滤波行人跟踪算

2021-08-05法付丽梅

软件工程 2021年7期

法付丽梅

摘  要:针对城市中智能视频監控问题,提出一种基于选定区域RGB直方图的粒子滤波行人跟踪算法。首先实现了Eiji Ota在2011 年实现的算法,该算法只能对红色目标进行跟踪,然后对该算法进行改进,改进后能够根据选定区域RGB颜色直方图的粒子滤波算法对行人进行跟踪。自动提取跟踪视频的第一帧,然后在第一帧上选择要跟踪的区域,再计算选定区域的RGB平均值,最后根据该区域的RGB颜色空间直方图进行粒子滤波行人跟踪。实验结果表明,采用改进的方法对视频监控的行人进行跟踪,效果较好。

关键词:RGB;直方图;粒子滤波;行人跟踪

中图分类号:TP391.41     文献标识码:A

Particle Filter Pedestrian Tracking Algorithm based on

Color Histogram of Selected Area

FU Limei

(Department of Software Engineering, Dalian Neusoft University of Information, Dalian 116023, China)

fulimei@neusoft.edu.cn

Abstract: Aiming at the problem of intelligent video surveillance in cities, this paper proposes a pedestrian tracking algorithm based on RGB (Red, Green, Blue) histogram particle filter in selected area. The paper first implements the algorithm implemented by Eiji Ota in 2011, which can only track red targets, and then improves the algorithm. After the improvement, it can track pedestrians based on the particle filter algorithm of the RGB color histogram of the selected area. The first frame of the tracking video is automatically extracted, and then the area to be tracked is selected on the first frame. After that, the RGB average value of the selected area is calculated, and finally, the particle filter pedestrian tracking is performed according to the RGB color space histogram of the area. Experimental results show that the improved algorithm is effective in tracking pedestrians under video surveillance.

Keywords: RGB; histogram; particle filter; pedestrian tracking

1   引言(Introduction)

视频监控技术就是利用计算机视觉来代替人工处理一些图像及视频信息,为人类减轻负担。采用智能监控系统可以利用计算机视觉技术代替人工进行视频的观察甚至分析。本文所研究的行人跟踪属于计算机视觉领域中的目标跟踪问题,可以在智能监控中进行应用,诸如安全、运营、交通运输等方面。

国内外一些城市部署了众多视频监控系统,对监控视频进行实时分析,可以发现其中的犯罪行为,从而发现并制止犯罪行为的发生,为防止犯罪行为和犯罪行为的侦破作出了贡献。对行人进行跟踪有更广泛的应用,可以为监控者减轻负担,也可以挖掘出诸如禁区侵入检测、打斗检测、区域人群密度检测等一系列分析,这些都具有重要的实用和经济价值[1]。同时,行人跟踪作为行人行为分析之前的步骤,也具有重要的研究意义,如果要对行人的行为进行分析就要持续并高效地对行人目标进行跟踪。……

登录APP查看全文