顾及多空间相似性的地下管线数据匹配

综合地下管线与专业地下管线是同一管线地物的两种表达形式。综合地下管线数据精准、概括;专业地下管线数据表达范围广、属性信息翔实。本文以天然气管线数据为例,选取管点关联管段分布形态作为衡量依据计算管线结构相似性,从管点本体概念名称和属性两方面计算管线语义相似性,以两管点间管段分布形态为特征计算管线形状相似性。以此空间相似性构建管点特征向量,采用SVM(support vector machine)支持向量机的分类方法及管点唯一匹配原则实现管点实体匹配。试验表明该算法能够有效解决管点匹配问题。

Underground Pipeline Data Matching Considering Multiple Spatial Similarities

Integrated and professional underground pipeline data are two forms of pipeline.Integrated underground pipeline data is accurate and general, while professional underground pipeline data expresses and contains detailed attribute information.Taking the data of natural gas pipeline as an example, this paper calculates structural similarity measured by the distribution pattern of pipelines that pipeline-point connects with, semantic similarity presented by the names and attributes of the pipeline-point ontology concept, and shape similarity characterized by the shape of arcs between two pipeline-points. The matching of pipe points is realized with the method of support vector machine classification algorithm and unique-matching principle combined with these spatial similarity. Test results show the matching of pipe points is well solved by the proposed algorithm.