The Query Techniques Of Spatial Database On R-tree

Spatial database technique is a hotspot in the area of database currently, and the research fruits have begun to be applied to many various fields.Spatial data have the characteristics of huge quantity, the high correlation between attribute data and spatial data, and the different data types with the different reflective geographical entities. As the huge quantity of spatial data and high complexity of spatial objects and spatial query, the efficiency of spatial query is one of the important factors evaluating the performance of spatial database. At the same time, practical applications have put forward impending requirements for the query efficiency of spatial database.In this paper, the basic techniques of spatial database query are firstly described, And then he researches from R-tree spatial index technique on spatial clustering, spatial toplogy-direction join-index based on R-tree, to the spatial join refinement for unevenly distributed spatial objects are emphasized, at last, on the basis of above researches, spatial data query experimenting system on GIS is designed and implemented.In the research work of R-tree index technology, the current algorithms of R-tree are firstly summarized, and then based on these algorithms, the hybrid spatial clustering algorithm is put forward and the corresponding R-trees are implemented in both dynamic and static environment. At last, some performance comparision tests are carried out and the performance of HCR-tree is proved.In the research work of spatial topology-direction join-index based on R-tree, the join relations of spatial objects and their estimating rules are introduced firstly, and then the concept of spatial join-index and its implementation methods on R-tree are described, on these basis, the concept and the constructing method of topology-direction join-index are put forword. The building process of spatial join-index is acturally the filter step of spatial join query processing. The building of topology-direction join-index on R-tree is almost the same with theprocess of common spatial join-index, it adds some topological and directional constraints in the process which are emphasized in this paper. In the end, to improve the flexibility and complexity of spatial join-index, the distance constraints are discussed.The spatial join processing on asymmetrical spatial objects is mainly relative to the refinement step of spatial join query processing. In the precondition of both datasets having built spatial join-index, according to the asymmetry characteristic of practical pages, the three-step processing strategy on asymmetrical spatial objects, which are clusters partitioning, partitions scheduling and pages scheduling, is proposed. The cluster partitioning is acturally the partitioning question of graph and the partition scheduling is the TSP question of graph in nature. In this paper, genetic algorithm is introduced to solve these problems such as cluster partitioning and partition scheduling. On the basis and results of above, some decisive rules of page accesses are provided. At last, some experiments and comparisons are carried out and the feasibility and advancement of this algorithm are proved.In the end, to validate and test the above algorithms, a new spatial data query experimenting system on GIS is designed and developed. The system is mainly on the basis of analysis of current spatial data query softwares and the application of research results of this paper. The main functions of this system include basic query, point query, region query, neighbor-nearest query and join query among which join query includes spatial topology query, direction query, distance query and any combination of these queries.