基于车牌照数据的通勤特征车辆识别研究

近几年来，交通拥堵日益成为大中城市最严重的交通问题之一，而由通勤行为 引起的早晚高峰交通拥堵最为突出，严重影响了城市居民的出行和交通系统的运行.因 此，本文从车辆的角度出发，采用上海快速路牌照识别系统采集数据，通过k-means 聚类 数据挖掘方法，提取路网中的通勤特征车辆，并分析了通勤特征车辆在路网中的出行时 空分布.分析得出，在上海快速路网中，占全部检测车辆2.8%的通勤特征车辆在早晚高峰 时提供了高达36%的交通量.在早高峰时段，识别出的通勤特征车辆主要分布路段为中环 外圈、陆家嘴方向、逸仙高架和沪闵高架；晚高峰时段，基本集中在与早高峰的相反方向. 研究结果表明，本文基于车牌照数据的数据挖掘方法，可以有效地提取通勤特征车辆并 研究其出行行为的时空特征，能够为城市交通拥堵问题的缓解和交通需求管理政策的提 出提供辅助决策信息.

Recognition of Vehicles with Commuting Property Using License Plate Data

Recently, traffic congestion is becoming one of the most traffic problem of the urban transportation. The traffic problem in rush hours caused by the commuting behavior is too prominent to bring about serious dilemma to the individual’s travel and the operation of transportation system. This paper aims to extract the vehicles with commuting property (WCVs) based on the collected vehicle license plate data by the vehicle license plate recognition system (VLPRS) in Shanghai, and analyzes the temporospatial distribution of the commuting vehicles. K-means clustering method is applied to data excavation. The results show that the vehicles with commuting property, which accounts 2.8 percent of all vehicles detective in VLPRS, supplied as the highest as 36% of traffic in peak hours. In morning rush hours, the main roads where WCVs travel are outer of the Middle Ring Line, Yan’an Elevated Road, Yixian Elevated Road and the Humin Elevated Road, while on the opposite direction in afternoon rush hours. It is proved out that the vehicles with commuting property can be effectively extracted on the basic of vehicle license plate data, and the analysis of travel behavior can provide a solid foundation for the more effective and meticulous transportation demand management, as well as the resolution of traffic congestion problem.