Research Of Spatial Join And Query Optimization Based Raster

The data is an important part of Geographic Information System. From the opinion of data structure, the raster data structure is a description of spatial geographic information, the simplest structure, also known as the grid structure (raster or grid cell) or the pixel structure (pixel). In fact, the raster data structure is a pixel array, each pixel fix its position with its rank. In the structure of raster, the point is expressed with a grid cell; the line is expressed with a group of neighboring grid cell; the surface or region is expressed with a collection of grid cells with the adjacent properties. Any objects (land usage, soil type, undulating terrain, environmental pollution, etc.) can be expressed with the raster data approximation. Although there are many spatial join algorithms based raster approximation, cost estimation and query optimization of spatial join operation need further study.Based on analyzing some spatial-temporal database query methods, this thesis proposed a new raster approximation suitable to performing spatial joins as a filter in the second step of the Multi-Step Query Processor, involving these two common data types (polyline and point) and the classes of different spatial joins involving them. We proposed a raster signature named—Four-Color Directional Raster Signature (4CDRS). Actually, that is to divide the object into sections and determine the type of the sections intersecting with the cells. When the direction of the section intersecting with the cell is horizontal, it is represented with Horizontal. When the direction is vertical, it is represented with Vertical. When they are not intersecting with each other, it is represented with Empty. It is represented with Inconclusive when other else possibilities. Calculating 4CDRS, we only need to determine the type of the section intersecting with the cell. Thus, each type can be quickly calculated, and we can only use 4CDRS to express polylines.The 4CDRS signature has the main advantage of faster generation time and that it can be used to represent polylines. We analyzed the type of the cells of the 4CDRS, provided an algorithm to generate 4CDRS. In the end, this thesis gave an algorithm for spatial joins query optimization which is based on the 4CDRS. Besides, we analyzed the algorithm, did the experiment and the results demonstrated the effectiveness of the approach.