线实体候选匹配集的优化方法研究

Methodological research on optimal matching candidates of line objects

地图数据合并是地理空间数据集成的基本途径,同名实体匹配是其中的重点与难点。本文根据线实体的形状将其分为简单线实体与复杂线实体,提出针对前者以线实体端点与中点为发生元生成的Voronoi图所得到的邻近对应关系为依据、针对后者以基于线实体缓冲区重叠度构造的相似性测度指标为依据优化候选匹配集的思路,并设计与实现了相关算法。实证研究表明,该算法能够适应不同比例尺与不同时相的城市道路网地图数据同名实体匹配。

As a fundamental way to integrate geospatial data, feature matching is the focus and hardship of map conflation. The article divided line objects into simple ones and complicated ones based on their shape, and presented an idea to optimize their matching candidates according to different criteria. For the former, the criterion was the proximal relationship of their endpoints and midpoints derived from Voronoi diagram ; for the latter, the criterion was a similarity index derived from overlapping proportion of their buffers. The corresponding algorithm was also designed and implemented. The case study testified that the algorithm could meet the needs of feature matching in urban road maps with different scale and phase.