基于共享单车轨迹的精细路网更新方法

城市道路数据的完整性和实时性是保障位置服务和规划导航路径的关键支撑。该文提出一种基于共享单车轨迹数据的新增自行车骑行道路自动检测和更新方法:首先,结合缓冲区方法和轨迹—路网几何特征检测增量轨迹;其次,基于分段—聚类—聚合策略提取更新路段,利用多特征融合密度聚类算法与最小外包矩形骨架线法提取增量道路中心线;最后,基于拓扑规则完成道路更新。以广州市共享单车轨迹为例,将该方法与传统栅格细化法进行实验对比,结果表明:该方法能有效更新道路网络,且在2 m和5 m精细尺度范围内提取的新增道路覆盖精度提升14%左右;在7 m尺度下精度达90%以上,在10 m尺度下精度达96%以上。

A Fine-Grained Road Network Update Method Based on Shared Bike Trajectory

The integrity and real-time of urban road data is the key support to guarantee location services and navigation path planning.Widely distributed and time-sensitive crowdsourced trajectory data provide new ideas for road data generation and updating.However,most of the existing automatic road extraction methods are based on vehicle trajectories for global road network topology extraction,with low sampling frequency and insufficient road coverage,which makes it difficult to meet the practical needs of diverse navigation location services.In view of the above,we propose an automatic detection and update method for newly added roads based on shared bike trajectory data.Firstly,the incremental change trajectory is detected by combining buffer method and trajectory-road network geometric features.Secondly,based on the segmentation-clustering-aggregation strategy to extract the updated road sections,a multi-feature fusion density clustering algorithm containing midpoint distance threshold,angle threshold,and length threshold is proposed for sub-trajectory clustering,and then the incremental roads are extracted based on the trajectory clusters using the minimum bounding rectangle(MBR)skeleton method.Finally,the road update is completed based on the topological rules.The proposed method is compared with the traditional raster refinement method using real shared bike trajectories in Guangzhou.The experimental results show that the method can not only effectively update the road network,but also improve the accuracy of the new road coverage extracted by the proposed method by about 14%in the fine buffer range of 2 m and 5 m;in the buffer range of 7 m and 10 m,the extraction accuracy of the proposed method can reach more than 90%,especially in the scale of 10 m,the accuracy can reach more than 96%,which verifies the effectiveness and accuracy of the proposed method.