基于视频识别的互通合流区交通冲突空间分布  

Spatial Distribution of Traffic Conflicts in Interchange Merging Area Based on Video Recognition

作　　者：赵笑月 汪攀 朱顺应[3] 蒋若曦 邹禾 ZHAO Xiao-yue;WANG Pan;ZHU Shun-ying;JIANG Ruo-xi;ZOU He(Wuhan Transportation Development Strategy Institute,Wuhan Hubei 430017,China;Wuhan Planning and Design Co.,Ltd.,Wuhan Hubei 430014,China;School of Transportation,Wuhan University of Technology,Wuhan Hubei 430063,China)

机构地区：[1]武汉市交通发展战略研究院,湖北武汉430017 [2]武汉市规划设计有限公司,湖北武汉430014 [3]武汉理工大学交通学院,湖北武汉430063

出　　处：《公路交通科技》2021年第5期90-99,共10页Journal of Highway and Transportation Research and Development

基　　金：国家自然科学基金项目(71771183)。

摘　　要：为深入研究高速公路互通合流区交通冲突空间分布特征,克服传统数据采集的条件限制,弥补断面数据采集的缺点,利用无人机视频采集高精度、区域连续的多车辆轨迹数据,并运用OpenCV视频自动分析系统和Matlab扩展工具,依据帧间差分法和时空上下文视觉跟踪算法原理,进行车辆识别与跟踪,记录采集范围内车辆每一帧的微观车辆运动数据。运用交通冲突识别指标TDTC(Time Difference to Collision)同时识别同车道和异车道冲突,将交通冲突分为严重冲突、中度冲突和一般冲突,并使用累计频率法确定严重冲突与中度冲突的阈值为3.1 s,中度冲突与一般冲突的阈值为7.8 s。依据地面坐标绘制直观的交通冲突空间分布图,并进行了量化统计分析。发现在靠近互通合流区匝道出口处及合流区末端是交通冲突比较集中且交通冲突比较严重的区段;整个互通合流区的严重冲突率呈现先增后减再增加的趋势;主线交通量比匝道交通量对交通冲突率空间分布的影响大,匝道大车比例比主线大车比例对交通冲突率空间分布的影响大。In order to deeply study the spatial distribution characteristics of traffic conflicts in expressway interchange areas,overcome the limitations of traditional data acquisition and remedy the shortcomings of cross-section data acquisition,by using the high-precision and continuous multi-vehicle trajectory data collected by drone video as well as the OpenCV video automatic analysis system and MATLAB extension tool,according to the principles of inter-frame difference method and spatio-temporal contextual visual tracking algorithm,the vehicle recognition and tracking are carried out,and the microscopic vehicle motion data of each frame of the vehicle within the collection range are recorded.The conflicts in the same lane and different lanes are identified simultaneously by using the traffic conflict identification indicator TDTC(Time Difference to Collision).The traffic conflicts are divided into severe conflicts,moderate conflicts and general conflicts,the threshold value of severe conflict and moderate conflict is determined as 3.1 s and the threshold value of moderate conflict and general conflict determined as 7.8 s by cumulative frequency method.According to the ground coordinates,an intuitive spatial distribution diagram of traffic conflicts is drawn,and the quantitative statistical analysis is conducted.It is found that(1)the positions near the exit of the ramp in the interchange and the end of the merging area are sections where traffic conflicts are concentrated and more serious;(2)the serious conflict rate of the entire interchange merging area shows a trend of first increasing,then decreasing and then increasing;(3)the traffic volume of the main line has a greater influence on the spatial distribution of the traffic conflict rate than the traffic volume of the ramp,while the proportion of large vehicles on ramp has a greater influence on the spatial distribution of traffic conflict rate than the proportion of large vehicles on the main line.

关 键 词：交通安全 交通冲突 冲突空间分布 视频识别 冲突严重性 

分 类 号：U491.2+65[交通运输工程—交通运输规划与管理]