摘要
对人体目标识别和跟踪的研究是计算机视觉领域一个重要的方面,具有很大的实用价值,其目的是让计算机代替人类完成诸如视频监控和人流量统计等任务.根据公交车上下客的特点,作者构建了一个统计乘客人数的系统,并提出了一种基于差分统计的计数算法.该算法既可以独立完成乘客人数统计的任务,也可以作为具有更高准确率计数算法的前端算法.实验结果表明,该算法的计数准确率达到85%以上,具有实际的应用价值.
The research on human object recognition and tracking is an important part of computer vision, which is of great practical value. It aims at making the computer accomplish some missions instead of human, such as video supervision and pedestrian flow calculation. According to the characteristics of bus passengers, a system for calculating bus passengers has been constructed and a counting algorithm based on difference statistics has been proposed in this paper. This algorithm can accomplish the mission of pedestrian calculation independently, and also can be a front algorithm of the calculation algorithm which possesses higher accuracy. The experiment result has shown the calculation accuracy of the algorithm can reach more than 85 %, so it has practical application value.
出处
《四川大学学报（自然科学版）》
CAS
CSCD
北大核心
2007年第4期825-830,共6页
Journal of Sichuan University(Natural Science Edition)
关键词
人体目标
识别
跟踪
人流量
human object, recognition, tracking, pedestrian flow
