基于车牌识别数据的机动车出行轨迹提取算法

为了提取城市路网上所有运行车辆的出行轨迹，系统科学地再现所有车辆的运行场景，进而为分析城市交通需求的结构和时空分布特性提供数据支撑，提出基于车牌识别数据的机动车出行轨迹提取算法.通过车牌及时间戳排序提取出行链；利用相邻节点间的速度，结合交叉口邻接矩阵完成行链的分离；基于K则最短路径算法（KSP算法）及灰色关联法（GRA算法），对出行轨迹进行补全重构.对贵阳市南明区的实际车牌识别数据进行算法测试.结果表明，提出的基于车牌识别数据的机动车出行轨迹提取算法在测试区域的综合准确率大于92%.
Abstract: The vehicle trajectory extraction algorithm based on license plate recognition data, in order to extract the travel trajectory of all vehicles running on the urban road network, systematically reproduce the operation of all vehicles scene, and provide data support for analyzing the structure and spatial and temporal distribution characteristics of urban traffic demand. The travel chain through the license plate and timestamp was sorted out; The separation of the travel chain was accomplished by using the velocity between adjacent nodes and the intersection matrix; The travel trajectory based on K shortest path algorithm (KSP algorithm) and Gray relational analysis (GRA algorithm) was completed. The algorithm was tested based on the actual license plate identification data of Nan-ming District, Guiyang City, Results show that the algorithm based on the license plate recognition data is more than 92% in the test area.