一种有效的混合空间索引机制

Efficient Hybrid Structure of Spatial Index for GIS Databases

论文结合网格索引和R树索引的特点,提出了一种基于粗分网格和聚类R树的混合索引方案(HSIBGR)。该方案首先将矩形地理空间进行粗网格划分建立一级索引,然后针对每个小网格建立基于R树的二级索引,采用聚类技术处理溢出R树结点的分裂。论文还对HSIBGR索引机制进行了理论分析和实验验证,结果表明HSIBGR索引机制能有效地减少存储开销,提高检索效率,具有良好的性能。

According to the characteristics of spatial index of grids and R-tree,this paper proposes a hybrid structure of spatial index based on wide grids and clustering R-tree（HSIBGR）.In HSIBGR,a rectangular region of geography is roughly partitioned to several sub-regions,which have same areas and shapes.For objects in every sub-region,we construct spatial index of R-tree.Nodes splitting of R-tree are implemented based on clustering techniques.The authors analyze performances of HSIBGR though theoretical methods and evaluate those by experimental results.The results show that HSIBGR can effectively reduce overheads of storages and improve efficiency of searching.