交通视频分析与跟踪系统设计
Design of Traffic Video Analysis and Tracking System
DOI: 10.12677/JISP.2018.74027, PDF,  被引量    科研立项经费支持
作者: 赵 洁, 刘 波:天津城建大学计算机与信息工程学院,天津;王光夫, 韦彦光, 孙 杰:天津瑟威兰斯科技有限公司研发部,天津
关键词: 车辆检测车辆跟踪车流量统计车辆压线检测Vehicle Detection Vehicle Tracking Vehicle Flow Statistics Violation Detection of Traffic Line
摘要: 交通视频分析与跟踪系统主要涉及运动车辆检测与跟踪、车流量统计、车辆压线检测三大部分。本文主要利用背景差分法和大灯提取算法分别获取日间和夜间视频中的运动车辆前景。在计数算法中,计算前景目标和大灯目标的质心坐标,并设置虚拟计数脉冲矩形,当目标质心点触发两个脉冲则进行计数。在压线车辆检测算法中,使用基于HSV颜色空间和Hough变换直线检测的算法获取道路内部的黄线区域,将运动车辆轨迹中点的坐标大于等于黄线上点的坐标作为判断压线的依据。本文以OpenCV开源库和Python编程语言为开发工具,实现了包含运动车辆检测、车辆跟踪、车流量统计和车辆压线检测的交通视频分析与跟踪系统。
Abstract: The system of traffic video analysis and tracking includes three major components: vehicle detec-tion and tracking, vehicle flow statistics and violation detection of traffic line. In this paper, back-ground difference method and headlight extraction algorithm were used to obtain the foreground of moving vehicles in video at day and night respectively. In counting algorithm, the centroid co-ordinates of the foreground and headlight targets were calculated and the virtual count pulse rec-tangle was set. The vehicle was counted when the target centroid point triggered two pulses. In the algorithm for violation detection of traffic line, the yellow line area inside the road was detected based on HSV color space and the Hough transform for line detection. The violation of traffic line is judged when the midpoint coordinate of moving vehicle’s trajectory was greater than or equal to the coordinates of points in the yellow line. In this paper, the system of traffic video analysis and tracking was implemented consisting of moving vehicle detection, vehicle tracking, vehicle flow statistics and violation detection of traffic line based on OpenCV open-source library and Python development tools.
文章引用:赵洁, 刘波, 王光夫, 韦彦光, 孙杰. 交通视频分析与跟踪系统设计[J]. 图像与信号处理, 2018, 7(4): 236-248. https://doi.org/10.12677/JISP.2018.74027

参考文献

[1] 王锶佳, 周妍. 中国智能交通现状调查[J]. 交通企业管理, 2017, 32(4): 12-14.
[2] 王彬. 基于视频的车流量检测算法研究[D]: [硕士学位论文]. 兰州: 西北师范大学, 2015.
[3] 冯仕民. 智能视频监控中运动目标跟踪技术研究[D]: [硕士学位论文]. 成都: 电子科技大学, 2010.
[4] 周航. 运动目标检测及跟踪技术研究[D]: [硕士学位论文]. 重庆: 重庆大学, 2013.
[5] 高韬, 刘正光, 张军. 基于交通视频序列的多运动目标跟踪算法[J]. 中南大学学报(自然科学版), 2010, 41(3): 1028-1036.
[6] 王菽裕, 刘瑞芝, 闻江, 等. 基于车头灯对的夜间车辆检测[J]. 电视技术, 2016, 40(8): 112-117.
[7] 董雪雪. 夜间车辆检测算法研究及其应用[D]: [硕士学位论文]. 南京: 南京信息工程大学, 2017.
[8] Du, X. and Tan, K. (2016) Vision-Based Approach towards Lane Line Detection and Vehicle Localization. Machine Vision and Applications, 27, 175-191.
[Google Scholar] [CrossRef