Abstract: Spatial index is the key technology to realize spatial query, and its performance decides the storage efficiency of spatial data and the performance of spatial query directly. In order to improve spatial query efficiency, a hybrid spatial index structure called LQR-tree is proposed. For the existing QR- tree index structure, there may be a problem that smaller objects fall into larger nodes in the allocation of nodes. The structure will combine loose quardtree and R-tree and can move nodes from the upper level to the lower level. The index can optimize query  performance of moving spatial objects. In this paper, the structure of LQR-tree is given and insert and delete algorithm are put forward. And relevant theorems and proofs are given.

Key words: spatial index, spatial database, spatial query, R-tree, loose quardtree

摘要： 空间索引是实现空间查询的关键技术，其性能的好坏直接决定着空间数据的存储效率及空间查询的性能。为了提高空间查询效率，提出一种混合空间索引结构松散QR-树：LQR-tree。针对已有的QR-树索引结构在节点分配中，可能存在较小的对象落入较大的节点中的问题，将松散四叉树和R-树相结合，能够实现节点下移，优化处理移动空间对象的查询，给出LQR-tree的结构和插入删除算法，并提出对应算法的相关定理和证明。

关键词: 空间索引, 空间数据库, 空间查询, R-树, 松散四叉树