| Starting with analysing the basic features of spatial data, this paper points out that the property and spatio-temporal characteristics are the basic elements of spacial data and that they can not be segmented. By analysing the limitations of spatio-temporal data management of traditional GIS, this paper sets forth the realibility and necessity of developing the Temporal Geographic Information System (TGIS). At the same time, this paper also discusses the history, application backgroud and the future trend of TGIS.Because the establishment of spatio-temporal data model is the core of TGIS, this paper discusses the recent study and development of spatio-temporal data model in the world and comments on the characteristics and the problems of the existing spatio-temporal data model. As the raster data is an very important kind of data structure of GIS and its matrix data structure is very suitable for describing the natural phenomenon which change continuously, this paper points out the significance and necessity of developing the raster data spatio-temporal data model from the point of such application fields as landscape change simulation, city dynamic evolution, fire spread and land desertification expanding.(1) According to the thoughts of Base State with Amendments, for the first time this paper points out the Raster Data Spatio-Temporal Data Model (RD-STDM) which is based on the quaternary linear quadtree encoding compression (MQ Codes), and discribes RD-STDM's conceptual model, logical model and phisical model.(2) By comparing and analysing several commonly used compression methods for raster data, this paper decides to use quaternary linear quadtree encoding (MQ codes) compression algorithm. And the author proposes and accompolishes the thoughts and the algorithm of Fast MQ codes compression based on the relational database and the stack structure for the first time, this will lay a good and necessary basis for the management of spatio-temporal raster data under the RD-STDM.(3) In order to further reduce the storage of the spatio-temporal raster data, this paper proposes and analyses sex types of spatio-temporal variations of raster data based on the Run-Length Encoding compresstion and three types of spatio-temporal variations of rasterdata based on the MQ codes compression for the first time. And this paper also gives the thoughts of how to realize the algorithm on the computer. So this can help us to manage the spatio-temporal raster data easily according to the thoughts of RS-STDM. Obsviously, this also lays a good basis for futher spatio-temporal ayalysis and history backdating.(4) According to the set theory, the author gives the thoughts of how to overlap the diversity of the raster spatio-temporal changing based on the MQ codes compression for the first time. And analyses the four cases of overlapping the diversity of the raster spatio-temporal changing during the operation. Moreover, the author also designs the algorithm and gives the program. So this will lay a basis for calculating the snapshot at a certain given time.(5) Based on the thoughts of RD-STDM, the algorithm of MQ codes compression, the algorithm of finding the diversity of raster spatio-temporal changing and the algorithm of overlapping the raster spatio-temporal changing, the author designs and realizes a spatio-temporal raster database under the SQL Server 2000 system. And the author develops an prototype of raster spatio-temporal data management system on the basis of the raster database. And tests and verifies the thoughts of RD-STDM and realated algorithms by using the the raster data of land-use of Yanchi County during 1989 and 2003. The results show that all the thoughts of RD-STDM and related algorithms are correct and attain the prospective target. It is very competent for the RD-STDM to store, manage, analyse and output the spatio-temporal raster data.(6) Lastly, this paper also discusses the thought of integrating the RD-STDM, GIS and CA model.
|
PDF Full Text Request