摘要
首先,依据摄像机的高度和角度提出了一种加权面积的计算方法;然后,将加权面积作为主要特征,利用人体头部和躯干部位梯度方向的不同,结合AdaBoost分类器对人群密度进行智能分级;最后,依据密度分级的结果对不同密度级别的人群分别进行人数统计。实验结果表明,所提出的方法能有效、实时、准确地对地铁站场景的人群进行人数统计。
Crowd counting in dense crowd scene has become an important and difficult topic in the field of automatic surveillance system.Since there are some requirements in the orbital traffic,a crowd counting approach for underground railway station is proposed.A weighted area computing approach is presented according to the information of camera firstly.Then the crowd density is estimated using AdaBoost classifier with the difference of gradient orientation between head and body of people.The people number is counted according to the crowd density and the weighted area.The experiment results show that the proposed approach is effective and feasible for crowd counting in underground railway station.
出处
《武汉大学学报（信息科学版）》
EI
CSCD
北大核心
2012年第3期343-347,共5页
Geomatics and Information Science of Wuhan University
基金
中央高校基本科研业务费专项资金资助项目(6081001)
关键词
地铁站台
加权面积
密度分级
人数统计
underground railway station
weighted area
density estimation
crowd counting
