基于速度的空间轨迹停留点提取算法  

Algorithm Study for Stay Points Recognition of Spatial Trajectory Based on Velocity

作　　者：侯颖超 王盼成[1] 刘兴权[1] 滕洁[1] 

机构地区：[1]中南大学地球科学与信息物理学院地理信息系,湖南长沙410083

出　　处：《地理与地理信息科学》2016年第6期63-68,共6页Geography and Geo-Information Science

基　　金：湖南省自然科学基金项目(10JJ6061)

摘　　要：空间轨迹中的停留点提取是将空间轨迹转换到语义轨迹的关键步骤。该文将速度变量引入停留点的提取,提出基于速度的时间聚类算法和速度聚类算法解决现有方法中的"伪停留点"和停留点丢失的问题。基于速度的时间聚类算法首先沿时间轴将轨迹点进行聚类得到候选停留点,然后利用速度阈值过滤候选停留点,得出实际停留点。速度聚类算法首先通过对速度的判断选取候选停留点,然后根据空间距离阈值对候选停留点的空间距离进行过滤,得出实际停留点,解决了停留点判断中的漏判问题。实验表明,基于速度的时间聚类算法对出租车轨迹数据(稳定时间间隔、不存在长时间轨迹点缺失)的空间轨迹停留点识别效果较好,而速度聚类算法更适用于步行轨迹(可能存在长时间轨迹点缺失)的分析。Stay point recognition is a key step to convert spatial trajectory to semantic trajectory.Velocity is employed to recognize stay points.Velocity-based time clustering algorithm and velocity clustering algorithm are presented to solve pseudo stay points and losing of stay points.Given time threshold,trajectory points are clustered by velocity-based time clustering algorithm along the time axis to get candidate stay points,then real stay points are filtered out by given velocity threshold.To avoid missing stay points,the velocity clustering algorithm divides the stay point recognition task into two phases:1)Stay point candidates are obtained through the judgment of velocity;2)Real stay points are filtered out from stay point candidates by the given spatial distance threshold.Experiments show that velocity-based time clustering algorithm can recognize stay points of the spatial trajectory with stable sampling time interval well,while velocity clustering algorithm can be better adapted to the pedestrian track points which sometimes may disappear for a long time.

关 键 词：停留点 空间轨迹 速度聚类 时间聚类 

分 类 号：P208[天文地球—地图制图学与地理信息工程]