摘要
自然邻近关系在空间数据库管理系统中具有十分重要的地位。为了克服Voronoi图受空间目标位置的影响,本文采用约束Delaunay三角网,针对在二维欧式空间中离散面目标的自然邻近关系查询问题。首先对离散面进行三角剖分,然后将代表相同邻近关系的三角形合并成一个union,进而建立面目标自然邻近关系表达结构-Unions Delaunay。通过与Unions Delaunay相结合建立空间索引结构-R-tree Gridfile来实现对空间目标自然邻居完备的候选集提取以及索引结构的更新。对比实验结果表明R-tree Gridfile索引结构能够有效支持空间目标的自然邻近关系查询。
Natural neighbor relationship plays an important role in spatial database management system. In order to overcome the influence of spatial target location on Voronoi diagram, the problem of natural adjacency query for discrete areal objects in two-dimensional Euclidean space is discussed. The method uses the constrained Delaunay triangulation, firstly, the discrete surface is triangulated, and then the triangles that represent the same adjacency relationship are merged into a union, and the surface adjacency relationship structure that is depicted by Unions Delaunay is established. By combining with Unions Delaunay, the spatial index structure-R-tree Gridfile is used for accomplishing the complete candidate set extraction of the natural neighbor of space target and updating its index structure. The comparative experiment results show that the R-tree Gridfile index structure can effectively support the natural adjacency query of spatial objects.
出处
《地理信息世界》
2018年第1期32-38,共7页
Geomatics World
基金
国家自然科学基金(41561082
41161061)资助