ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume X-3/W2-2022
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., X-3/W2-2022, 87–92, 2022
https://doi.org/10.5194/isprs-annals-X-3-W2-2022-87-2022
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., X-3/W2-2022, 87–92, 2022
https://doi.org/10.5194/isprs-annals-X-3-W2-2022-87-2022
 
27 Oct 2022
27 Oct 2022

POI POINT ENTITY MATCHING AND FUSION BASED ON MULTI SIMILARITY CALCULATION

J. Zhao1,2,3, X. Niu1,3, Y. Cui1,3, Y. Zhao4, M. Guo1,3, and R. Zhang1,3 J. Zhao et al.
  • 1School of Geomatics and Urban Spatial Informatics, Beijing University of Civil Engineering and Architecture, Beijing, China
  • 2Key Laboratory for Urban Spatial Information of the Ministry of Natural Resources, Beijing 102616, China
  • 3Beijing Key Laboratory for Architectural Heritage Fine Reconstruction & Health Monitoring, Beijing 102616, China
  • 4Guangzhou Panyu Polytechnic, Guangzhou, China

Keywords: POI, Multi-source fusion, Text Similarity, Distance Similarity, Big Data Update, Analytic Hierarchy Process

Abstract. This paper presents a multi-source POI matching method with multi feature similarity, which can effectively solve the problem of low matching accuracy of POI data from different sources. The spherical distance method, editing distance method and Jaro Winkler method are combined to calculate the distance, name, address distance and other main attributes of POI data. Then the importance of each feature index is analyzed by using analytic hierarchy process, and the feature weight of each similarity is obtained. The candidate matching objects are screened according to the total similarity to determine the final matching object. Finally, POI points are fused by selecting spherical center coordinates, name aliasing and address normalization methods. Experiments show that the recall and accuracy of this method for POI matching point recognition are significantly higher than those based on name similarity and distance similarity. The recall rate increased by 17.43% and 5.17% respectively, and the accuracy rate increased by 4.37% and 1.22%.It provides more comprehensive and accurate data support for urban function analysis and smart city construction.