摘要
针对商业区分析中涉及商业区边界划分和商业区中心的判断的问题,该文提出了一种基于互联网兴趣点(POI)数据的商业区自动提取方法。首先,基于核密度函数,利用POI计算研究区域的商业网点密度;其次,使用主体密度模型的特征点分类方法寻找潜在的商业中心,并基于POI所在位置的商业密度,使用自然断点法筛选出属于商业区的POI;最后根据商业区POI的平均最近邻距离,将聚集的点构建Delaunay三角网,提取商业区。以北京市市辖区为例,对百度地图上与商业相关的POI进行商业区探测挖掘,研究结果表明,基于互联网POI数据得到的北京市商业区比根据阈值从核密度图提取等值线的方法更加符合真实商业区情况,可用于城市商业分析与规划。
Commercial area analysis involves delimitation of commercial area and identifying of commercial centers,and this study proposes a method to dig out urban commercial area information automatically based on points of interest（POI）from the Internet.First of all,commercial dots density in study area is calculated with POI by kernel density function;secondly,potential commercial centers are found by the method of feature point classification on density model,and natural breaks is used to filter POI belong to commercial area based on density value;finally,Delaunay triangulation is used to delimit commercial area according to average nearest neighbor distance.Taking downtown of Beijing city as study area,POIs from Baidu map are used to analyze commercial area,and the results show that the method in this study delimit commercial area more appropriate to compare with the method of delimiting by contours,this study result could be used to analyze and plan the urban commerce.
作者
童莹
郭庆胜
王勇
TONG Ying;GUO Qingsheng;WANG Yong(School of Resouce and Environmental Sciences,Wuhan University,Wuhan 430070,China;Chinese Academy of Surveying and Mapping,Beijing 100036,China)
出处
《测绘科学》
CSCD
北大核心
2018年第11期57-62,78,共7页
Science of Surveying and Mapping
基金
国家自然科学基金项目(41471384)
国家重点研发计划项目(2017YFB0503502
2017YFB0503601)
中国测绘科学研究院基本科研业务费项目(7771705)
关键词
POI分析
核密度估计
空间识别
商业区
POI analysis
kernel density estimation
spatial recognition
commercial area
