Spatial Indexing Algorithm Based On Fractal Theory

Fractal (Fractal) theory is a new branch of modern mathematics.Fractal geometry is a non-regular geometric shapes of the geometry. As the anomalies is common in nature, therefore, also known as fractal geometry of nature described. Through in-depth study of fractal theory to prove some properties of the Peano curve, especially for Hilbert Space-Filling Curves, the index for space research provides the necessary theoretical knowledge.Spatial database of spatial information technology has been a core research component of the field. As the spatial information infrastructure and spatial data access technology, rapid development, the increasing scale of spatial data, spatial data sharing become increasingly demanding, at the same time, spatial data warehouse, spatial data mining on spatial database system performance in the growing demand. Relying on hardware to improve performance of database systems increasingly difficult circumstances, to improve the spatial data sharing capabilities, and enhance the efficiency of spatial data indexing become a hot research frontier.Based on fractal theory, by generating a Hilbert curve, space efficient and rational division of the data, and spatial indexing system with the current widely used R-tree spatial index method, the formation of a new spatial indexing method and system, a good solution spatial indexing speed and accuracy of indexing problems, and effectively improve the spatial data distributed mass index efficiency of space. As follows:In-depth study of the fractal image coding theory, L systems and iterated function system fractal graphics rendering methods, and gives the formation of Hilbert Space-Filling Curves program, scan matrix algorithm designed to accurately encoded by Hilbert;After analyzing the impact of the spatial locality, a spatial data declustering method based on Hilbert curve hierarchical decomposition was presented In this method, the adjacent relations among objects are kept by Hilbert curve. According to the unstructured variable-length and non-uniform distribution characteristics of spatial data, smaller size grid is used in the area where spatial objects are very dense while lager size grid is used in the area where spatial objects are very sparse. The balanced distribution of spatial data among computer nodes can be achieved by hierarchical decomposition of initial partitioning grid;A new spatial index structure (HR-tree index). Clustering properties using Hilbert curve to solve the R-tree search path of non-uniqueness problem; by the minimum bounding rectangle decomposition (DMBR) to maximize the compression index of the redundant data generated to obtain a more accurate search results.Experiments show that the fractal spatial index algorithm to do the research, which can effectively overcome the current lack of spatial data distribution method, in the process of spatial index on a relational database through a direct scale of operation and maintenance between, as far as possible to avoid calling the system relies on the ArcGIS ArcObjects package under development platform correlation function, the function of spatial index while significantly improved the spatial index efficiency, spatial data applications provide critical spatial index system and a good support for data sharing.