多尺度道路网的距离匹配算法研究

Abstract

根据道路网折线的匹配特点,提出基于格网索引的折线——结点距离匹配算法,将复杂的折线与折线之间的几何相似度计算转换为求结点到折线距离的匹配方法,降低了计算复杂度,并通过建立格网索引来提高计算效率。在应用实例中,采用曲线拟合的最小二乘法确定算法的匹配容差和匹配成功率之间的关系,并通过与现有统计匹配算法比较,可知该算法效率高且匹配成功率较理想,能够满足多尺度道路网数据匹配的应用需求。 Existing algorithms for geometry matching of multiple entities are short of meeting the requirements of multi-scale road network analysis due to algorithms’ high complexity and long processing time.Based on the analysis of matching characteristics of poly-lines on road networks,this paper suggests the geometric similarities between poly-lines can be evaluated using the distances between poly-lines and nodes.As an alternative to a direct match of geometric similarities between poly-lines,this line-node distance-based algorithm has a much lower computational complexity and the potential for efficiency improvement with a grid-based spatial index.Through a case study,we compared the distance-based algorithm with the traditional matching algorithm in computational efficiency and matching quality.We found the proposed algorithm is superior in both domains and will be able to meet the requirements for multi-scale road network analysis.We also conducted an analysis on the relationships between matching tolerance and precision using the least square method.