地理研究

地理研究 ›› 2013, Vol. 32 ›› Issue (3): 452-464.doi: 10.11821/yj2013030007

• 城市与社区研究 • 上一篇    下一篇

青岛城市道路邻近中心性及其应用方法

王瑶莉, 高松, 刘瑜   

  1. 北京大学遥感与地理信息系统研究所, 北京100871
  • 收稿日期:2012-01-17 修回日期:2012-08-30 出版日期:2013-03-10 发布日期:2013-03-10
  • 通讯作者: 刘瑜(1971- ),男,副教授,博士。主要从事地理信息科学研究。E-mail:liuyu@urban.pku.edu.cn
  • 作者简介:王瑶莉(1990- ),女,硕士。主要从事地理信息科学应用研究。E-mail:wangyaoli0810@yahoo.com.cn
  • 基金资助:

    国家高技术研究发展计划项目(2011AA120301);国家自然科学基金项目(40801046,40928001,41171296)

Exploration into urban street closeness centrality and its application methods：A case study of Qingdao

WANG Yaoli, GAO Song, LIU Yu   

  1. Institute of Remote Sensing and Geographic Information System, Peking University, Beijing 100871, China
  • Received:2012-01-17 Revised:2012-08-30 Online:2013-03-10 Published:2013-03-10

PDF (PC)

赞

被引次数

18 | 4

摘要： 以青岛为实证研究区域,从理论角度探讨了道路全局和局部邻近中心性的地理含义及其各自优缺点,认为全局邻近中心性(GCC)反映的是网络上各点到网络“质心”的距离,可从全局角度反映区位重要性;局部邻近中心性(LCC)反映小范围空间区域的重心,与局部路网密度相关。进一步根据核密度估计(KDE)的机理证明,在城市地理空间结构的实际应用案例中适宜采用KDE法来进行空间分析。本文以手机话务量数据反映的城市居民活动空间结构为例讨论了邻近中心性核密度估计(CC-KDE)的应用,发现与居民活动相关性最高的是核密度全局邻近中心性(GCC-KDE),因为全局邻近中心性核密度估计(GCC-KDE)综合了到网络“质心”距离与道路网络局部密度双重因素,故反映城市人口等与道路密度相关的社会经济现象更有效。

关键词: 城市道路网络结构, 邻近中心性, 核密度估计, 居民活动分布, 手机话务量, 青岛

Abstract: Using Qingdao as a case study and mainly focusing on methodology,this paper investigates the geographic implication,advantages,and disadvantages of global and local street network closeness centrality.The global closeness centrality(GCC) can reflect the distance to the centroid of a network,and indicate the importance of a location on a global scale.Meanwhile,the local closeness centrality(LCC) can be used to find the mass center of relatively small ranges,which is corresponding to the street network density.Further study finds that the kernel density estimation(KDE) of closeness centrality combines both the distance to the centroid on the whole and the street density on a local scale,and thus shows a stronger potential to represent urban struture,such as the human activity intensity of a city.This research introduces mobile phone Erlang value,an indicator of communication intensity,to estimate human activity intensity and uses it to examine the geographical impification of KDE closeness centrality.We find that the KDE of global closeness centrality(GCC-KDE) has the strongest positive correlation with human activity distribution,since GCC-KDE takes into account both the distance to the mass center of a network and the local street denstiy,and thus it can be a protential indicator of socio-economic phenomenon correlated with street density.

Key words: urban street network, closeness centrality, kernel density estimation, distribution of human activities, mobile Erlang value, Qingdao