基于图文法的城市道路匹配方法  

A Node Matching Method of Urban Road Networks Based on Graph Grammar

作　　者：包永钢 BAO Yonggang(Taiyuan Institute of Surveying and Mapping,Taiyuan 030001,China)

机构地区：[1]太原市测绘研究院,山西太原030001

出　　处：《地理信息世界》2020年第5期102-107,共6页Geomatics World

摘　　要：提取多源城市路网交叉口节点,每个节点都有唯一的结构模式信息,将城市道路网的节点匹配转化为图匹配中的最大公共子图问题,采用最小图编辑距离衡量两份道路网数据中道路节点的匹配程度,并且加入距离和方向统计,寻找结构形态最为相似的道路节点。实验表明,相对于传统的点线匹配方法,当两份数据存在不均匀偏移时,基于图论的匹配方法可以更完全地利用拓扑信息,使用上下文相关的拓扑分析方法帮助提高几何和语义的匹配结果,该方法不依赖于任何非空间语义信息,原则上可以应用于各种数据源,有效提高匹配准确度。This study extracted the nodes of multi-source road networks data,each road intersection has the unique structure semantic information,then realizes the node matching of urban road networks into by maximum common subgraph for graph matching,which using minimum edit distance to measure the matching degree of two road node data and distance and direction statistical elements to find the most similar structure of road nodes.The result shows:compared with traditional approaches,when the two raw data have uneven deviation,this approach is able to better utilize structural information by using the context of topological analysis so as to improve the efficiency and accuracy of geometry matching results.As non-spatial information is not needed in this method,this method can be applied to various data sources.

关 键 词：图文法 城市路网 结构模式 节点匹配 

分 类 号：P208[天文地球—地图制图学与地理信息工程]