POI点要素空间邻近图质量评价——以Delaunay三角网和缓冲区构建方法为例  

Quality Evaluation on Spatial Proximity Graph of POI Based on Delaunay Triangulation Network and Buffer Construction Method

作　　者：苏友能 徐青[1] 朱新铭 邹霓霄 武钰清 刘波[2] SU Youneng;XU Qing;ZHU Xinming;ZUO Nixiao;WU Yuqing;LIU Bo(Geospatial Information Institute,Information Engineering University,Zhengzhou 45000l;School of Surveying and Geoinformation Engineering,East China University of Technology,Nanchang 330013,China)

机构地区：[1]信息工程大学地理空间信息学院,河南郑州450001 [2]东华理工大学测绘与空间信息工程学院,江西南昌330013

出　　处：《地理与地理信息科学》2024年第3期14-20,共7页Geography and Geo-Information Science

基　　金：国家自然科学基金项目(42161064);国家自然科学基金青年项目(42101454);江西省自然科学基金项目(20232ACB204032)。

摘　　要：为有效促进地理空间数据邻近关系的快速构建与语义表达,基于结构、数量、空间、性能等多重视角,提出结合图结构分析、度频率分析、可视分析、推理性能分析的空间邻近图评价方法,并利用Delaunay三角网和缓冲区两种方法进行上海市交通服务设施POI点要素空间邻近图构建与质量评价。结果表明,相比缓冲区方法,基于Delaunay三角网构建的空间邻近图具有如下优势:①在结构方面,规模小、图上节点聚集程度低、冗余邻近关系少;②在数量方面,邻近关系数量少且集中、局部邻近关系冗余度低;③在空间方面,邻近关系数量与POI空间分布情况相关性强;在性能方面,推理效率稳定。该方法能对空间邻近图进行全面评价,可为地理空间知识图谱构建过程中邻近关系建立方法的选择提供参考。In order to effectively promote the rapid construction and semantic expression of geospatial data proximity relation-ship,the quality evaluation method of spatial proximity graph is studied in this paper.Firstly,based on multiple perspectives such as structure,quantity,space and performance,a spatial proximity graph evaluation method combining graph structure anal-ysis,degree frequency analysis,visual analysis and reasoning performance analysis has been proposed.Then,taking Delaunay tri-angulation network and buffer as two construction methods,the paper constructs and evaluates the quality of the spatial proxim-ity graph of POI of Shanghai traffic service facilities.The results show that compared with the buffer method,the spatial prox-imity graph constructed by Delaunay triangulation network has the following characteristics:In terms of structure,it has the characteristics of small scale,low clustering degree of nodes on the graph,and fewer redundant proximity relationships on the whole.@In terms of quantity,it has the characteristics of fewer and concentrated proximity relationships and low redundancy of local proximity relationships.③In termsofspace,thereisastrongcorrelation between thenumber ofproximity relationships and the spatial distribution of POI.@In terms of reasoning performance,it has the characteristics of stable reasoning efficiency.The evaluation method proposed in the paper will comprehensively evaluate the spatial proximity graph,and provide reference for the selection of the method of establishing proximity relationships in the process of constructing the geospatial knowledge graph,and help with the expression of geographic knowledge and the improvement of service level.

关 键 词：邻近关系 空间邻近图 DELAUNAY 缓冲区 POI 知识图谱 

分 类 号：P208[天文地球—地图制图学与地理信息工程]