Settlement Matching Algorithm Using Spatial Similarity Relations as Constraints

Settlement matching is one of the kernel parts of multi-source spatial data fusion and multi-scale data updating. Following the cognition habits of mankind in finding strange buildings, the spatial relationship similarity is used to assist the settlement matching process. The discrete computing method according with human cognitive habits is proposed after analyzed the similarity of topological relationship, distance relationship and direction relationship. And the matching processes are as fellows. Firstly, the outstanding settlement of the original object is picked up and computed to find its matching object. Secondly, referencing the matched object, the next matching object is achieved by the extend-first traversal to unmatched objects. Thirdly, the precision matching is fulfilled by traversing every settlement object all in this way. Finally, the matching quality is evaluated by comparing the spatial relationship similarity of adjacent objects. Test illustrates that this method can effectively improve the matching precision in the case of data hardly displacement and high settlement shape homogeneity.