摘要： QR-树处理海量空间数据时，其深度和R-树内目录矩形的重叠面积会变大，导致查询效率降低。针对该问题采用K-means算法对索引对象进行聚类分析，构造新的聚类中心使其能处理具有多种形体的索引对象，并在QR-树中引入超结点存储聚类结果。提出一种QCR-树空间索引结构来提高查询效率，给出QCR-树的插入、删除和查询算法。实验结果表明QCR-树的查询性能优于QR-树，适用于海量数据。

关键词: 空间索引, QR-树, QCR-树, K-means算法, 超结点

Abstract: The depth of QR-tree and the overlapping areas of directory rectangles of R-tree will increase when the massive spatial data is processed by the QR-tree, which incures lower query efficiency. Aiming at this problem, this paper carries out clustering analysis of index objects by K-means algorithm, and a novel formula of clustering center is constructed to make K-means deal with index objects with various forms. It introduces super nodes for storing the clustering results and proposes a QCR-tree spatial index structure to improve the query efficiency. The insertion, deletion and query algorithms of QCR-tree are presented. Experimental results show that QCR-tree, whose query performance is higher than QR-tree, is fit for processing the massive data.

Key words: spatial index, QR-tree, QCR-tree, K-means algorithm, super node

中图分类号: 

• TP311.12