基于AI的交通流状态智能检测与拥堵态势分析  

Intelligent detection of traffic flow state and congestion situation analysis based on AI

作　　者：相阳 陈晓静[2] Xiang Yang;Chen Xiaojing(Jiangsu Communications Construction Bureau,Nanjing 210019,China;ITSSKY Technology Co.,Ltd.,Nanjing 210036,China)

机构地区：[1]江苏省交通工程建设局,江苏南京210019 [2]江苏长天智远交通科技有限公司,江苏南京210036

出　　处：《江苏科技信息》2023年第15期54-57,61,共5页Jiangsu Science and Technology Information

摘　　要：文章基于YOLO+DeepSORT框架,开发了一套基于AI的交通流状态智能检测系统,能够快速准确地检测出不同地点的交通流量。文章以润扬大桥附近区域为例,进行交通流特性分析,并且从交通流演化规律出发,可以得到产生拥堵的3种可能原因,包括高速公路车道减少、匝道汇入车流影响和下游扬州市区拥堵等。这些结论对未来的高速公路大流量管控工作具有重要的参考意义。The article is based on the YOLO+DeepSORT framework and develops an AI based intelligent detection system for traffic flow status,which can quickly and accurately detect traffic flow at different locations.Then,taking the area near the Runyang Bridge as an example,traffic flow characteristics are analyzed,and starting from the evolution law of traffic flow,three possible causes of congestion can be identified,including the reduction of highway lanes,the impact of ramp merging traffic flow,and downstream congestion in Yangzhou urban area.These conclusions have important reference significance for future highway traffic control work.

关 键 词：高速公路 交通拥堵 视频检测 交通流特征 

分 类 号：U4[交通运输工程—道路与铁道工程]