过刊检索
年份
《城市交通》杂志
双语出版
20190102-Data Mining and Urban Road Traffic Analysis Based on Traffic Camera Data
点击量:956

Wang Bei1, Ning Pinghua1, Duan Xiaomei1,Wang Shiming2, SzetoWai Yuen3
(1.Guangzhou Municipal Engineering Design & Research Institute Co. Ltd., Guangzhou Guangdong 510000, China; 2.Guangzhou Public Security Bureau Traffic Police Detachment, Guangzhou Guangdong 510000, China; 3.The University of Hong Kong, Hong Kong 999077, China)

Abstract: Based on traffic data collected by cameras, this paper introduces the traffic analysis method for roadway design and data mining algorithms, and explores their applications in urban transportation planning, infrastructure construction, and traffic management at macro, mesa, and micro levels. Based on the spatial and temporal sequences of vehicles, three analysis tasks and solution methods are proposed to obtain three important indexes, namely the traffic flow distribution among adjacent areas, vehicle trajectories, and the time-dependent traffic operations. The concept of traffic flow distribution among adjacent areas is proposed. It refers to the traffic volumes between two adjacent areas connected by roads, and it reveals the macroscopic distribution feature of traffic flows. The application of frequent sequence mining algorithms can identify the sequences of vehicle clusters passing by the data collection cameras, which reveals vehicle clusters travel areas and paths. Based on the vehicles entering and egressing time, the travel time by hour of the day can be calculated. Finally, taking Yichang city in Hubei province as a case study, the paper illustrates the data mining and road traffic analysis framework based on traffic camera data.

Keywords: traffic engineering; traffic camera data mining; vehicle license plate matching; traffic flow distribution among adjacent areas; vehicle trajectories; time-dependent road operational level