基于聚类的AOI出入口位置提取方法研究  

Cluster-Based Extraction Method of Entrances and Exits in AOI

作　　者：钱振 周侗[1] 陶菲[1] 范紫灵 刘浩 陆杰 QIAN Zhen;ZHOU Tong;TAO Fei;FAN Zi-ling;LIU Hao;LU Jie(School of Geography,Nantong University,Nantong 226007,China)

机构地区：[1]南通大学地理科学学院,江苏南通226007

出　　处：《地理与地理信息科学》2020年第2期1-7,F0003,共8页Geography and Geo-Information Science

基　　金：国家自然科学基金项目(41301514);南通市空间信息技术研发与应用重点实验室项目(CP12016005);江苏省大学生创新创业训练计划项目(201910304113Y);国家级大学生创新创业训练项目(201910304036Z)。

摘　　要：AOI出入口是行人寻址与车辆导航的重要地标,城市化导致出入口位置和通行状态频繁变动,通告不及时则会影响正常出行,出入口的更新问题亟须一种自动化的探测方案。由于出入口与客流的聚集点存在一定的正相关,而出租车的下客点在一定程度上可以代表周边客流量,该文据此提出了一种基于出租车下客点的最高热度出入口位置提取方法,以城市的AOI和POI作为地理背景数据,以出租车轨迹和路网作为出行行为数据,通过改进的地理位置文本聚类获取AOI位置,通过层次聚类并结合二分K均值聚类提取下客点簇,将两者进行关联计算得到出入口的具体位置。该方法既可应用于出入口位置的快速提取,也可作为出入口变动的动态监测方案。As the important landmarks for pedestrian addressing and vehicle navigation,entrances and exits in AOI are changed frequently in their location and traffic status due to urbanization.Daily travelling will be affected without the timely updating of notification.So the update of the entrances and exits need to be automatically detected.There is a positive correlation between the entrances and exits and the passenger flow which can be represented by the taxi drop-off points.Based on this,this paper proposes a method to extract the location of entrance and exit by taxi drop-off point data.This method takes the AOI and POI of the city as the geographic background data,the taxi trajectory and road network as the travel behavior data.The improved geographic location text clustering is used to get the AOI location,and hierarchical clustering combined with bisecting K-means clustering are used to extract drop-off points cluster,then the locations of the entrances and exits can be calculated by correlation analysis of the AOI location and drop-off points cluster.This method can be used to quickly extract the entrance and exit positions and monitor the changes of entrances and exits.

关 键 词：聚类 AOI 时空数据挖掘 车辆轨迹 

分 类 号：P208[天文地球—地图制图学与地理信息工程]