Research Of Spatial Data Index For Power GIS System

With the continuous development of global information, there are more and more application of geographic information system(GIS), such as the power GIS system. Today’s GIS system, not only need to deal with the data in two-dimensional space, but also to deal with the data in three-dimensional or higher dimensional space. Therefore it needs to increase the requirements of spatial database. A key technology to limit the performance of spatial database is the spatial index, its performance is an important indicator to measure the overall performance of the GIS. The study of the spatial index becomes one of the priorities for the development of GIS.The data of power system has its particularity. For example, there are more point and line objects and fewer polygon objects, and some individual line objects may across a large range of space. The traditional grid indexing mechanism will increase the storage overhead of the system, and the object-based index will cause distortion of the line object which affect query efficiency. Therefore, the power system needs to explore a suitable index structure.This paper reviewed the research background and did a brief introduction of direction development of the spatial index. On the basis of the organization for spatial data of the traditional GIS technology, combined with the characteristics of data in the power GIS system. The paper put forward a organization for spatial data suited the power GIS system.The paper proposed a two-stage hybrid index mechanism on the basis of the advantages of the two types of index structures. The idea is creates the first index in accordance with the grid index, each grid contains a number of objects. Then establishes the second index for the objects which are entirely contained in the same grid. At last establishes indexed list for the objects that exist across the grid phenomenon. Through the two-stage index structure to improve overall system’s performance.The second indexing mechanism Hilbert R-tree has been improved by introduced on the basis of cluster thinking. It improves the clustering of spatial objects in the leaf nodes and reaches the purpose to improve the performance of the index mechanism. And the operating algorithms are designed for improved Hilbert R-tree and hybrid index structure.Finally, the paper proposed the theoretical analysis and comparative experiments for the performance of the hybrid indexing mechanism. The experiments verify the superiority of the hybrid indexing mechanism.