Research And Application On Spatial Database Index Technology Based On R-tree

With the rapid development of computer technology, in particularly the development ofInternet, Geographic Information System (GIS) emerged at a historic moment. In GIS,Spatial data as a basic element of a map is the core of the GIS, and spatial database is acollection for storage, management and operation of spatial data. Spatial database indextechnology is a core technology in spatial database and GIS. According to the characteristicsof spatial data, developing reasonable and efficient approach has become the research focusof spatial database and GIS. Spatial database index technology is the key to improve theretrieval performance of spatial database. So using spatial index is very necessary, spatialdatabase index technology has a very important significance in spatial query andconstruction of spatial database.This paper summarizes and analyzes the defects and advantages of the various spatialdatabase index technologies, the main research is index algorithm based on R-tree. As to thecharacteristics of R-tree, when the target quantity of index space is increased, the indexspace of intermediate nodes overlap and increase rapidly, thus, the search path will increaseand the efficiency of index will degrade rapidly. This paper proposes a modified R-treeindex structure: SR-R-tree, SR-R-tree is composed of a SR-tree and n R-tree, SR-tree(Spatial Relation Index tree) is used to represent spatial relations of geospace; R-tree is usedto store spatial objects. The SR-R-tree is connected by corresponding pointer of nodes. Thenthis paper gives the algorithm description of creation, searching, insertion and deletion ofSR-R-tree. The experiments prove that the modified R-tree index structure can properlysolve some problems of R-tree index algorithm in some ways of application. Compared withR-tree, SR-R-tree has more occupation storage space and better search performance. Themore index objects, the more obvious advantages of overall performance of SR-R-tree, butinsertion performance is worse than R-tree. At last, this paper applies SR-R-tree indexstructure to route-select service system, to make the path searching time being reducedgreatly and play a good performance.