基于视频的交通流信息采集算法的研究-《计算机技术与发展》

文章信息/Info

Title:: Research on Traffic Flow Information Collection Algorithm Based on Video

Author(s):: LIANG Feng-lan ; QIN Chuan; SHI Hua-ji; [1]Department of Computer, College of Suqian[2]College of Computer, Jiangsu University

Keywords:: traffic information collection; virtual loops detection; background difference; inter-frame difference

摘要:: 为了解决高速公路的交通流参数的检测过程中车辆跨道行驶、阴影和遮挡物影响而导致的车辆漏检或误检等问题，文中通过在交通视频图像中感兴趣的区域设置“T”形虚拟线圈，同时在分析了车辆通过虚拟线圈过程中YUV颜色空间中Y分量的变化趋势的基础上，设计了一种帧间差分与背景差分相结合的车辆检测算法，并：芷此基础上完成车速的测量。实验证明文中的算法能在一定程度上解决车辆跨道行驶时重复计数、由于车辆阴影或其他遮挡导致的车辆漏检或误检等问题

Abstract:: In order to solve the problems such as inter-road vehicles, missed or false detections of the vehicles caused by the shadows or occlusion in the detection process about the highway traffic flow parameters,it designs an algorithm of combining inter-frame difference and background difference and completes measuring the speed by setting up T-shaped virtual loops of the Jinterested region in the traffic video and analyzing the transformation trend of Y-component in YUV color space when the vehicles gets through the virtual loops. The experiment shows that the algorithm can solve the problems in a certain degree of duplicate counting, missed or false detection of the ve- hicles caused by the shadows or occlusion in the detection process