三维GIS中混合树空间索引结构的研究

Research on Spatial Index Structure of Hybrid Tree In 3D GIS

目的在空间数据库应用中,解决随着数据量的增加,单一的空间索引结构制约检索性能的问题.方法提出一种三维GIS中基于八叉树和R*树的混合树空间索引结构OR*树,该结构在对三维索引空间进行八叉划分的基础上应用R*树索引技术,将操作空间限定在某一特定的区域.结果实现了R*树以及OR*树的试验程序,并通过均匀分布的随机数据测试了不同层数OR*树的空间开销与插入、删除、查找所访问的磁盘页数.结论试验数据分析表明,该方法有效降低了R*树的插入、删除与查询操作的开销,数据量越大OR*树的效率越高.

In spatial database application, to solve the problem that a single special indexing structure constrains retrieval performance with the increment of the amount of data, a hybrid tree special indexing structure is proposed, and so is OR ＊ -tree, which is based on Octree and R ＊ -tree irl 3D GIS. It partitions the index space into eight subspaces in each level first, then creates a R ＊ -tree structure for every index space. The experimental procedure of R ＊ -tree and OR ＊ -tree is accomplished, and the experiment has tested the special costs of different lever number OR ＊ -tree and the accessed disk page number in the process of insertion, deletion and searching through uniformity distribution random data. The result of experiment indicates that the method reduces efficiently the operation costs of insertion, deletion and searching for R ＊ -tree. The greater the amount of data is, the better the efficiency is.