摘要
车辆轨迹中停留点提取是分析车辆用户语义行为的关键步骤.该文深刻剖析车辆轨迹特征,提出基于语义行为的车辆轨迹停留点提取算法来解决行车中遇到十字路口、立交桥等形成的"伪停留点"问题和无法捕捉短暂停留的停留点丢失问题.该算法首先通过切割系数的选取实现对停留稳定性的判断,进而聚类得到候选停留点,解决停留点丢失问题;然后利用车辆轨迹在空间路网分布极不均匀的特征,对其进行路网建模实现十字路口的拓扑模型,并利用距离阈值的选取过滤候选停留点,解决停留点中的误判问题.出租车历史GPS轨迹数据的实验结果表明,算法在召回率和准确率上相比现有轨迹停留点提取算法ETC和OC有一定的稳健性,同时在调和平均值上有明显提高.
The stay point extraction in the vehicle trajectory is a key step in analyzing the semantic behav⁃ior of the vehicle user.This paper deeply analyzes the characteristics of vehicle trajectories,proposes an algorithm based on the semantic behavior of vehicle trajectory stay points to solve the problem of"pseudo stay points"caused by crossroads,overpasses etc,and the problem of missing stop points that cannot be captured for a short stay.The algorithm firstly determines the stay stability by choosing the cutting coeffi⁃cient,and then clusters the candidate staypoints to solve the problem of stay point loss.Then the vehicle track is used to distribute the non-uniform features of the spatial network,and the spatial road network is modeled to realize the topological model of the intersection,and the distance threshold is used to filter the candidate stop points to solve the problem of misjudgment in the stay point.The experimental results of historical GPS trajectory data of the taxis show that the proposed algorithm has a certain degree of robust⁃ness in recall rate and accuracy compared with the existing trajectory stay point extraction algorithms ETC and OC.At the same time,it has a significant improvement in the average value of harmonic adjustment.
作者
陶健
王睿
秦晓安
殷西祥
TAO Jian;WANG Rui;QIN Xiao-an;YIN Xi-xiang(Department of Electronic Information Engineering,Anhui Business College of Vocational Technology,Wuhu 241002,China)
出处
《通化师范学院学报》
2021年第2期95-102,共8页
Journal of Tonghua Normal University
基金
安徽省重大线上教学改革研究项目(2020zdxsjg123)
院级自然科学一般项目(2020KYZ01)
安徽省教学教研项目(2019jyxm0714)
安徽省高校优秀青年人才支持计划重点项目(gxyqZD2020056)
院级三平台两基地重点项目(2020ZDX05)
院级三平台两基地重点项目(2020ZDX05)。
关键词
车辆轨迹
停留点
语义行为
停留稳定性
十字路口
vehicle trajectory
stay points
semantic behavior
stay stability
crossroads
