Comparison of Polygon Matching Methods in Multi-represented Vector Datasets

The Geospatial data production from different sources with varied precision and scales has experienced an extensive growth by increase of geospatial data usage in daily life. Through these diverse datasets in some of the applications, there is a need to identify corresponding features by geospatial data matching problems. In other words, the goal of geospatial data matching is to identify corresponding features which are the identical entities in real world, but they may be represented by different geometries, scales, precisions, and spatial relations in different datasets. The use of geospatial data matching problems in different studies like data conflation or authoritative data enrichment demonstrates the significant importance of such these problems in different fields of geospatial science. In general, feature matching problems can be classified into point matching, linear matching, and polygon matching based on the type of spatial feature applied in the matching process. Most of the previous studies have focused on linear matching. Therefore, this study is based on the polygon features which are more complex in comparison with the linear features, and most of recent studies have focused on this class. In this paper, the polygon matching problems have been investigated in three aspects of similarity measures, matching algorithms, and corresponding relations by comprehensive and precise study of previous work. Furthermore, the classifications in each aspect and the advantages and disadvantages of different methods have been presented.