Exploring the construction and application of spatial scene knowledge graphs considering topological relations. Issue 3 (22nd February 2022)

• Record Type: Journal Article
• Authors: Shen, Jingwei; Shi, Kaifang; Ma, Mingguo
• Is Part Of: Transactions in GIS. Volume 26:Issue 3(2022)
• Page Start: 1531
• Page End: 1547
• Publication Date: 2022-02-22
• Subjects: Geographic information systems -- Periodicals; 910.285
• DOI: 10.1111/tgis.12911
• Languages: English
• ISSNs: 1361-1682

Abstract

Abstract: A spatial scene consists of a large number of geographic objects and their spatial relations and attributes. It is important to explore an effective method to organize spatial scenes so that they are more readable by humans and machines. The construction of a spatial scene knowledge graph considering topological relations, which can play important roles in spatial query, spatial analysis, and spatial data modeling, is explored in this article. To construct the spatial scene knowledge graph considering topological relations, the definitions of the spatial scene knowledge graph schema, topological relation knowledge acquisition and storage, and spatial scene knowledge graph visualization are explored. Two typical applications, including knowledge reasoning and spatial neighbor analysis based on spatial scene knowledge graphs, are presented. A case study is designed, and the construction and application of a spatial scene knowledge graph of the Three Gorges Reservoir Region is shown. Nine contains relations, 18 crosses relations, and 47 touches relations are marked; and topological relation reasoning and spatial neighbor relation analysis based on a successfully constructed knowledge graph are performed. The results show the following: (1) the spatial scene knowledge graph considering topological relations is an effective and intuitive way to express the topological relations between geographic objects; and (2) the spatial scene knowledge graph may play important roles in graph‐based application areas, such as topological relation reasoning and spatial neighbor analysis.