A New R-tree Space Index Based on the Cluster of Grid Density and Dynamic Grid Division

The cluster based on the density is a kind of cluster analysis, its main merit is to discover cluster of arbitrary shape and insensitive to the noise data. R-tree organizes the space index according to the spatial data, the spatial overlapping is in a big way, the inquiry efficiency is low, in order to adapt the need of massive GIS inquiry application, this paper firstly carries on the division of the data set according to the grid density, uses the dynamic grid to carry on the expansion of its periphery data objects and the cluster to the R-tree nodes, makes the neighboring data be divided in the same smallest surrounding rectangle as far as possible, reduces the area of MBR, thus this index can reduce the overlapping of region divided in R-tree, enhance the retrieval efficiency.