Construction And Application Of Spatial DataWarehouse

The database management patterns, especially data center pattern, achieved data integration, datashare and separated applications from data in petroleum exploration and exploitation.But the knowedpatterns can not complete data analyzing and spatial data mining. This paper provides a new kind ofSpatial Data Warehouse model based on the Analysis of demands of exploration and the datacharacters.The proposed logical model of Spatial Data Warehouse, which summarizes and absorbs theadvantages of data warehouse, 3D geo-modeling, 3D GIS and data management of oil-gas explorationand exploitation, fullfills the demand of geo-data analyzing.Then, the physical model of Spatial DataWarehouse and the Spatial OLAP method are studied aiming at the problems of geo-data integralanalyzing and the multi-scale characters of spatial data.1) Based on the concept of data warehouse multi-dimension analyzing, this paper presentsexploration and exploitation data multi-dimension analyzing theory based on data warehouse andconstruction patterns of spatial data warehouse.2) The primary problem of spatial data warehouse is the multi-scale of spatial data. The existingmulti-dimension models don't take spatial and spatial scale into account; consequently the models havenot capabilities of analyzing, processing and representing in space and scale, and are not fit for spatialmulti-dimension data cube model. This paper consider spatial scale and dimension together, spatial scaleis considered as a dimension, equally, dimension is considered as a general scale, a multi-dimensiondataset logical model of data warehouse based on the concept of scale and a pattern of multi-scalespatial aggregation are proposed.3) In order to construct a fact table to meet the demands of geo-analyzing and representmulti-dimension relationship of geo-spatial better, this paper proposed that the fact table should beconstructed based on the spatial relationship of data, so called spatial fact table,then a star or snowflakeform datacube can be constructed of which the core is spatial fact table and the dimensions aregeo-properties.4) An octtree subdivision voxel model which represents spatial multi-scale is proposed; multi-scalerepresenting of geology body and geology field and spatial integration are implemented; fact table isconstructed; the relationships of dimensions and between spatial fact and dimensions are analyzed;multiple order voxel code based on decimal Morton code is constructed to relate spatial to properties, sothe relationships between objects (dimensions) can be constructed through spatial relationship. Allabove form the foundation of multi-dimension analyzing.5) Three kinds of analysis methods are studied, including the methods of spatial analyzing basedmulti-dimension model such as neighborhood analyzing, spatial cut, virtual drill and so on; the methodsof spatial measurement such as distance, volume, surface area and so on, the methods ofmulti-dimension analyzing such as slicing dataset, refining dataset based on spatial fact table, OLAPand so on.6) A prototype of spatial data warehouse based on above theory is implemented and used to modelDong Ying depression utilizing geology-geophysical data, this model is proved valid and practical.