A New PR-Tree Index Based on Pyramid Division

Because the number of data objects is large in the spatial database, generally speaking, the cost of spatial database query is larger than that of the relational database, the traditional R-tree index structure allows the overlapping and coverage among the sibling nodes, resulting in the duplicate storage of the spatial objects and multiple query paths, the inquiry and search efficiency are low, in view of the advantages of the pyramid division, a new PR-tree index based on pyramid division is proposed in this paper, divide the data objects into the smallest area surrounding them through endowing each queried object a certain unique pyramid value, the experiments show that the new PR-tree index can effectively avoid the duplicate storage of the data objects, the path of query and search is single, so it can well realize the various arithmetic operations of the data objects.