基于POI要素空间聚集特征的集镇识别与边界提取  

Market Town Recognition and Boundary Extraction Based on Spatial Clustering Characteristics of POI Elements

作　　者：余思汗 张楠[1,2] 陈瑞 刘超 杨顺[1] 孙嘉欣 蔡保祥 YU Sihan;ZHANG Nan;CHEN Rui;LIU Chao;YANG Shun;SUN Jiaxing;CAI Baoxiang(Earthquake Agency of Ningxia Hui Autonomous Region,Yinchuan 750001,China;Institute of Geology,China Earthquake Administration,Beijing 100029,China;Natural Resources Information Center of Ningxia Hui Autonomous Region,Yinchuan 750000,China)

机构地区：[1]宁夏回族自治区地震局,宁夏银川750001 [2]中国地震局地质研究所,北京100029 [3]宁夏回族自治区自然资源信息中心,宁夏银川750000

出　　处：《地理空间信息》2025年第4期100-104,共5页Geospatial Information

基　　金：中国地震局地震应急青年重点任务(CEAEDEM20240212);宁夏自然科学基金资助项目(2023AAC03807、2022AAC03701);基于遥感和人工智能的承灾体信息获取技术创新团队项目(CX2023-4)。

摘　　要：为了高效率精确识别集镇、准确提取边界,为地震应急快速评估和地震应急救援提供精准数据支撑,利用核密度分析、加权叠加分析和空间关联分析,探索集镇的空间分布和边界范围。根据POI数据的覆盖广、聚集度特征明显的特点,针对集镇职能提出居民依赖性合理计算方法,实现对研究区内所有集镇的识别和边界提取。结果表明:利用POI能够识别研究区内的全部集镇,并成功提取集镇边界,在空间重叠度和面积精度方面,具有较好的一致性和准确性,可为集镇识别、边界提取提供新思路。In order to efficiently and accurately identify market towns and extract their boundaries,to provide accurate data support for rapid earthquake emergency evaluation and rescue,we used kernel density analysis,WSA and spatial correlation analysis to explore the spatial distribution and boundary size of some market towns.Based on the wide coverage and obvious clustering characteristics of POI,we proposed a calculation method for town residents' dependency for the functions of market towns,achieving the recognition and boundary extraction of all market towns in the research area.The results show that using POI can successfully identify and extract the boundaries of all market towns.This method has fair consistency and accuracy in terms of spatial overlap and area accuracy,providing new ideas for town recognition and boundary extraction.

关 键 词：集镇 POI数据 空间聚集 核密度分析 加权叠加分析 地震应急 

分 类 号：P208[天文地球—地图制图学与地理信息工程]