基于轨迹数据的道路客运班车停留站点位置提取方法

识别并提取道路客运班车停留站点的位置, 可为道路客运的客运站站址优化、定制出行乘降站点设置、出行信息服务等提供依据和支持, 然而当前获取班车停留站点位置的方法存在成本高、周期长的问题。通过分析道路客运班车停留轨迹数据的典型特征, 以班车轨迹数据为数据源, 基于DBSCAN算法检测位于停留站点的点簇进而提取停留站点位置。同时, 针对DBSCAN算法具有高时间复杂度的问题, 通过建立格网索引对算法进行了改进。基于京津冀区域的136条道路客运班线的班车轨迹数据进行了实证分析, 结果表明: 改进DBSCAN算法提高了算法执行效率, 平均执行时间减少了59.72%, 且所生成的班车停留站点数量与传统算法基本一致; 在提取得到的282个班车停留站点中, 256个为真实的班车停留站点, 班车停留站点提取的正确率为90.78%。 

A Method for Extracting Regular Bus Parking Stops of Road Passenger Transport Based on Trajectory Data

It is of great significance to identify and extract the locations of the parking stops of regular buses of road passenger transport, providing support for optimizing the station location, setting the stops for customized travel, and the travel information service in road passenger transport. However, the current methods to obtain the location of parking stops of regular buses have problems of high cost and long cycle. The DBSCAN algorithm is used to detect point clusters located within parking stops and extract the parking stop location of regular buses by analyzing the typical characteristics of the track data of parking stops and taking those as the data source. Meanwhile, the DBSCAN algorithm is improved by establishing a special spatial grid index to decrease time complexity. Based on the track data of 136 regular routes of passenger service in the Beijing-Tianjin-Hebei region, this paper makes an empirical analysis.The results show that the improved DBSCAN algorithm can improve the execution efficiency, with the average execution time reduced by 59.72%. The number of parking stops is consistent with those generated by the traditional algorithm. Among the 282 regular bus parking stops extracted, 256 are real regular bus-parking stops with an accuracy rate of parking stops extraction of 90.78%.