基于骨架化和蜘蛛编码的面状实体匹配方法研究

同名面实体在不同来源的地图数据库中的矛盾性表达致使相关领域的研究出现诸多瓶颈问题,由此实体匹配作为基础研究成为首要解决的关键问题,本文采用两阶段匹配策略实现面实体匹配,预匹配中通过面积相异度进行初步判定,终匹配时先对面状实体进行骨架化从而降低点和线的维度,再用基于蜘蛛编码的方法实现匹配。最终实现的匹配算法由大比例尺地图数据库中的房产图和地形图进行匹配实验,结果证明了该算法的匹配效率和准确度。

Surface entity matching based on skeletonization and spider coding

The problem of contradictory expression of surface entity with the same name but from different map database led that the research of many related areas come forth much bottleneck,hence entity matching as basic research become primary key issue.The paper used two-stage matching strategy to match surface entity.During the stage of pre-matching,a preliminary determination was got by the dissimilar degree of area.While in the stage of final matching,it firstly skeletonized the surface entity thus reduced to the dimensions of the point and line,and then realized matching by the method of spider coding.Finally,the algorithm was experimented by property map and topographic map from large scale map database,meanwhile the efficiency and accuracy were proved as well.