基于车辆跟踪轨迹的停车和逆行检测研究

发布时间：2018-01-15 23:31

本文关键词：基于车辆跟踪轨迹的停车和逆行检测研究　出处：《长安大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着国民经济的快速发展,智能交通也越来越受到人们的重视。作为智能交通的重要组成部分,交通事件检测对于缓解交通压力、减少交通事故有着很重要的作用。基于视频的交通事件检测技术,由于其显著的优点(安装简单、覆盖范围广、检测直观),越来越成为各国学者研究的重点。本文在前人研究的基础上,重点研究了车辆检测和车辆跟踪,并根据质心坐标绘制运动轨迹,之后利用车辆跟踪轨迹分别对违章停车检测和逆行检测进行了研究。首先,在运动车辆检测方面,通过比较常用的背景提取算法,采用多帧图像平均法提取视频背景。分析帧间差分法、光流法、背景差分法,考虑实际检测效果,采用彩色背景差分法对车辆进行检测,之后对检测图像进行二值分割,形态学处理和团块填充,采用HSV颜色空间对车辆阴影进行去除,这一系列操作使车辆目标能够完整的提取出来。其次,在运动车辆跟踪方面,在介绍了常用跟踪方法的基础上,采用卡尔曼滤波和车辆质心特征相结合的跟踪方法,对下一时刻车辆特征值进行预测,之后对车辆进行搜索匹配,实现跟踪。通过分析质心坐标,提取出车辆跟踪轨迹,并实时显示在监控画面上面。在以上研究的基础上,对违章停车和逆行进行检测识别。对于违章停车,采用质心距离变化和速度变化进行判定,当质心距离和车辆速度小于一定的阈值时,结合具体停车时间进行判定。对于逆行检测,当车辆行驶方向和道路规定方向相反时,车辆质心的Y轴坐标变化来进行判定。最后分别对违章停车算法和逆行算法进行了视频测试,取得了很好的效果。
[Abstract]:With the rapid development of national economy, people pay more and more attention to intelligent transportation. As an important part of intelligent transportation, traffic incident detection can relieve traffic pressure. It is very important to reduce traffic accidents. The video-based traffic incident detection technology has obvious advantages (simple installation, wide coverage, intuitive detection). On the basis of previous studies, this paper focuses on vehicle detection and vehicle tracking, and draws the motion track according to centroid coordinates. Then the vehicle tracking track is used to detect the illegal parking and the retrograde detection. Firstly, in the aspect of moving vehicle detection, the background extraction algorithm is compared. The multi-frame image averaging method is used to extract the video background, the inter-frame difference method, the optical flow method and the background difference method are analyzed. Considering the actual detection effect, the color background difference method is used to detect the vehicle. After the detection image binary segmentation, morphological processing and block filling, the use of HSV color space to remove the vehicle shadow, this series of operations make the vehicle targets can be extracted completely. Secondly. In the aspect of moving vehicle tracking, based on the introduction of common tracking methods, the Kalman filter and vehicle centroid feature tracking method are used to predict the vehicle feature value at the next moment. After the vehicle search matching to achieve tracking. Through the analysis of centroid coordinates to extract the tracking track of the vehicle and real-time display on the top of the monitoring screen. On the basis of the above research. For illegal parking, the change of centroid distance and the change of speed are adopted, when the centroid distance and vehicle speed are less than a certain threshold. For retrograde detection, when the driving direction of the vehicle is opposite to the specified direction of the road. The Y-axis coordinate of the center of mass of the vehicle is changed to judge. Finally, the algorithm of illegal parking and the retrograde algorithm are tested by video, and good results are obtained.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;TP391.41

【共引文献】