一种利用低频轨迹数据提取道路交叉口的方法

为实现从低频轨迹数据中提取城市道路交叉口，本文设计了一种基于数据预处理与聚类算法的道路交叉口精准识别方法。首先结合轨迹数据的特征，采用启发式滤波算法对原始数据进行清洗，剔除冗余点与异常点；然后依据车辆的运行规律，提出了一种分步式道路交叉口的提取算法，由此计算出疑似道路交叉口的特征点；最后利用层次密度聚类算法(HDBSCAN)对筛选过后的轨迹点进行聚类并提取质心，得到道路的交叉口，最终以成都市某日的出租车行驶轨迹为数据源，进行试验分析。结果表明，使用该算法提取交叉口，精确率达95.33%、召回率达82.11%、F值达88.46%,能有效且准确识别城市道路交叉口信息，在城市管理与交通规划中具有一定的应用价值。

A method of extracting road intersections using low frequency trajectory data

In order to extract urban road intersections using low-frequency trajectory data, this paper designs a precise identification method of road intersections based on data preprocessing and clustering algorithm. Firstly, combined with the characteristics of the trajectory data, a heuristic filtering algorithm is used to clean the original data and eliminate redundant points and abnormal points. Then, according to the running rules of vehicles, a step by step algorithm for extracting road intersections is proposed to calculate the characteristic points of suspected road intersections. Finally, hierarchical density clustering algorithm (HDBSCAN) is used to cluster the selected track points and extract the centroid, which is the intersection of the road. Based on the data source of taxi driving track in Chengdu, the results show that,the algorithm can extract the intersection with an accuracy of 95.33%, a recall of 82.11% and F value of 88.46%. It can effectively and accurately identify the urban road intersection information, and has a certain application value in urban administration and traffic planning.