基于局部空间熵的亚洲城市扩张模式识别.PDFVIP

基于局部空间熵的亚洲城市扩张模式识别 刘凌波，吴昊，彭正洪 摘要：城市扩张对城市的经济、社会、健康和产生重要的影响，城市的过度扩张所导致的城市蔓延更会影 响城市的可持续发展系统，城市蔓延区的识别是城市空间结构研究的重要内容。随着大数据环境的发展， 越来越多的研究倾向于使用更为复杂和多样的变量来识别城市蔓延，但其结果仍然存在不确定性。本研究 立足于芝加哥学派城市空间结构模式和城市经济学密度衰减模式的解读，归纳出基于栅格人口密度分布的 四种城市扩张模式，总结不同城市区域在人口密度和局部空间密度均衡性的聚类特征，提出以人口密度和 局部均衡性对城市空间单元聚类以识别城市扩张模式和蔓延区域的设想。研究基于移动搜索法、摩尔邻域 和香农熵构造了局部空间熵方法以计算局部空间均衡性，通过K-means聚类对案例城市栅格人口数据和局 部空间熵进行聚类。实验结果验证了以单一栅格人口密度识别城市扩张模式的设想，城市蔓延可由人口密 度和其局部的均衡度进行识别。 关键词：城市扩张，城市蔓延，K-means聚类，移动搜索法，局部空间熵 IdentifyingUrbanExpansionPatternofAsianCitiesBasedonLocal Spatial Entropy Abstract: Urban expansion has an important impact on the economy, society, health aspects of cities ，wherein urban sprawl caused by excessive urban expansion will weaken the citys sustainable system. Identification of urban sprawl area plays a significant role on urban studies. With the development of big data environments, although more and more research had tended to use more complex and diverse variables to identify urban sprawl, the results still remained uncertain. Based on the interpretation of urban spatial structure model of the Chicago school and density decay model in urban economics,this studypresents four patterns ofurban expansion based on the gridded population density, summarizes the corresponding features of different kinds of urban areas by population density andlocal spatial evennessand assumestoidentify urban expansionpatternsand sprawl areasby clustering with population density and local spatial evenness. The study proposes local spatial entropy method to calculate local spatial evenness based on Floating Catchment Area method, Moore neighborhood and Shannon Entropy, implements the clustering with K-means algorithm. The results verify the assumption of identifying urban expansion pattern merely with griddedpopulation density,urban sprawl can be discriminated by population density anditslocal spatial even