融合OSM路网与POI数据的城市功能区识别研究  

Research on Urban Functional Area Recognition Integrating OSM Road Network and POI Data

作　　者：丁彦文 许捍卫[1] 汪成昊 DING Yan-wen;XU Han-wei;WANG Cheng-hao(School of Earth Sciences and Engineering,Hohai University,Nanjing 211100,China)

机构地区：[1]河海大学地球科学与工程学院,江苏南京211100

出　　处：《地理与地理信息科学》2020年第4期57-63,共7页Geography and Geo-Information Science

基　　金：国家自然科学基金项目(41101374)。

摘　　要：识别城市功能区的分布特征可为城市空间结构及资源的合理配置等提供帮助。该文以南京市中心城区为研究对象,基于核密度估计算法提出了一种融合OSM路网与POI数据的城市功能区识别方法:利用OSM路网对城市进行自动分割,获得具有相同社会经济功能的研究单元;基于核密度估计算法实现POI设施的影响扩散,削弱了POI点的离散化现象,同时设置核密度带宽对比实验,分析带宽大小对功能区识别结果的影响。结果表明:较大带宽使POI影响范围扩大,功能区混合程度增高,识别精度较低;较小带宽使POI对邻近位置的影响难以扩散,POI自身分布对结果影响较大;而最适带宽组识别效果最好,精度达81.3%;研究区被分为22个功能区类型(6个单一功能区、15个混合功能区和1个综合功能区),识别结果与南京市实际情况相符,可为城市发展规划与管理提供借鉴。Recognizing the distribution characteristics of urban functional areas has a positive effect on the optimization of urban structure and the formulation of reasonable development planning measures.The traditional division of urban functional areas has the problems of high cost,poor timeliness and large errors.Taking downtown Nanjing as an example,this paper proposes a method for recognizing urban functional areas by integrating OSM road network and POI data.This method uses the OSM road network to automatically segment cities and obtain research units with the same socio-economic functions,making the urban block segmentation more reasonable.The analysis of the POI using the kernel density estimation method weakens the discrete phenomenon of the POI points,reduces the influence of noise,and realizes the influence diffusion of the POI in the neighboring locations.At the same time,the experiments of bandwidth comparison are set up to analyze the influence of bandwidth selection on the recognition results of functional areas.In addition,based on the traditional single functional area recognition,this paper comprehensively considers the complex functional structure of the city to recognize the urban mixed functional areas and the comprehensive functional areas.The results of the study show that the size of the bandwidth has a certain impact on the recognition type and accuracy of the functional area;the optimal bandwidth experimental group has the highest recognition accuracy,and the recognition results are consistent with the actual situation in Nanjing,provides a reference for urban spatial structure,urban planning,and rational allocation of urban resources.

关 键 词：城市功能区 OSM路网 百度POI 核密度估计 带宽实验 

分 类 号：TU984.113[建筑科学—城市规划与设计]