一种面向地理信息系统的空间索引方法

Abstract

空间检索技术是有效地管理和操纵空间数据的一种必要手段。本文分析了目前常用的空间检索方法在地理信息系统应用上的局限性 ,提出了一种基于凸多边形的空间索引技术。本文首先介绍了目前常用的几类空间索引技术及其特点 ,接着讨论了面向地理信息的空间索引技术面临的基本问题 ,提出了基于凸多边形的空间索引结构—— CP-树 ,并侧重分析了其空间运算算法及时空效率 ,最后指出了空间索引技术所存在的问题及下一步的研究方向。 It has been recognized in the past that the traditional Database Management Systems cannot handle efficiently spatial data in mutil dimensional space. This paper focuses on deriving efficient access methods for spatial objects in Geographical Information Systems (GIS). First, we discuss the classification of existing spatial indexing methods and point out the problems of them. Second, the approximate representation of spatial objects is given. Third, the structure and the algorithm of CP tree, which based on the convex polyhedra, are proposed in detail. Finally, we compare search performance between CP tree and R tree. It is found that CP tree requires a few CPU time and disk accesses for region searching than the R tree. Thus, it enhances the spatial query performance remarkably.