Abstract: A Pedestrian flow detection method in public transport system based on the computer vision is presented. By analyzing the video data of the spot pedestrian flow and using the techniques of the image processing, the pedestrian flow through the field of vision can be extracted from the video image. The algorithm of extraction of human head characteristic curve against a dynamic background low definition and specific shooting angle is presented. And the parameter estimating of ellipse template through partial ellipse is presented. The results show that the algorithm is correct and efficient.

Key words: Pedestrian flow detection, Computer vision, Edge extraction, Curve extraction, Head recognition

摘要： 提出了一种基于计算机视觉的公交系统人流量检测方法。通过分析现场行人视频数据，运用图像处理的方法提取图像中通行的行人以及通过视场内的人流量，阐述了复杂背景、低分辨率和特殊拍摄角度条件下的人头特征曲线提取算法和基于部分椭圆弧的椭圆模板参数估算法，试验结果表明了该算法的正确性和有效性。

关键词: 人流量检测, 计算机视觉, 边缘提取, 曲线提取, 人头识别

CLC Number: 

• TP391