Study On The Organization Method Of Spatio-temporal Data In GIS

Temporal GIS is the one of the important and difficult issues in GIS, which is not only the demand of the GIS theories development, but also the need of putting forward physically in GIS applications. The paper mainly studies the organization of spatio-temporal data that is the key in temporal GIS based on the relational database. And the main results are as the following:1. Put forward an integrated framework for organizing spatio-temporal data, which includes the integration of static state and temporal state, the integration of vector data and raster data, the integration of topological data and non-topological data, the integration of discrete change and continuous change, and the integration of geometry data and attributive data.2. Aiming at the vector spatio-temporal data, some concrete organization strategies are brought forward. Firstly, the precision of time is second and the time is represented with the number. Secondly, the discrete change is described with state and event, and the continuous change is described with event. Thirdly, the reality data and history data are separated, and moreover, the spatial data and attributive data are separated. Finally, the time is signed on entity and record, and non-topological data model is adopted to organize spatial data.3. Propose that the four spatio-temporal changes should be considered in organizing the vector spatio-temporal data, which include the discrete change of space and attribute, the continuous change of space and attribute, the discrete change of space and continuous change of attribute, and the continuous change of space and discrete change of attribute. Furthermore the geographic entity is classified to point, line and area, and the temporal layer of vector is divides twelve types. Finally, some relational tables of temporal layer are designed too.4. Based on the ideas of base state with amendments, the temporal layer of raster is designs by four relational tables, and the structure of tables is also given. Then the raster spatio-temporal data is organized based on two dimensional run-length, and three algorithms such as the compression, diversity analysis, and diversity overlapping are designed based on two dimensional run-length dynamic codes.5. An experiment system is designed and realized. Then some experiments are done with patch data, vehicle track data and cloud images of meteorological satellite. And the results show that the proposed designs and methods in the paper are efficient.