Research Of Vector Data Index Based On NAND Flash

Along with the development of information technology, computer technology, other high technologies and the development of new concept of future wars, information war has become the main battle form of future wars. The digital battlefield changes in a short time. In front of the mass information, a module which can analyze and support correct decision making to combat personnel in a short period of time to make plans is needed. The module is geographic information system.Data is the blood of GIS, and data management is the heart of GIS. Especially in the embedded environment, it is restricted by a hardware performance. Spatial data management is particularly important. The spatial data index is an important means of spatial data management. At present, spatial data index technology in PC environment has been very mature. R-tree and its change trees which are based on disk reading characteristics are adopted widely. On the foundation of the researches of R series index algorithms, considering the characters of vector data, CR*-tree is brought forward, which is more efficient and more adapted to the embedded NAND Flash. The main research contents are as follows:1. According to the development process of spatial index structure, R index algorithm and its change tress are analyzed. Based on this, the differences between spatial index in embedded environment and spatial index in PC environment are discussed. Then, three aspects that embedded spatial index can improve are put forward.2. The reading and writing attributes of NAND Flash is analyzed. In view of its characteristics, two MBR compression algorithms are put forward. They are relative coordinate method and relative coordinate - grid method.3. Considering that the efficiency of R*-tree index is relatively high and the algorithm is simple, the R*-tree is used in the proposed compression algorithm, then CR*-tree is proposed. Then, insertion, inquire, and deletion algorithms of CR*-tree are discussed. In terms of the completeness and system of index, log update way of embedded spatial index is put forward.4. Efficiency test experiments are designed. Node capacity, node accessing quantity and query time of R*-tree and CR*-tree are analyzed in theory. Then we test and verify the theories by vector data. The results show that the efficiency of CR*-tree is better than that of R*-tree.