Research On Spatiotemporal Raster Data Modeling Based On GRIC Model

With the value of geographic information has gradually been recognized, the GIS has got an unprecedented development. The amount of geographic information data is constantly expanding, and its applications in GIS are also rapidly enriched and increased. Due to the limitations of current acquisition method by space and time, the geographic information data has obviously discrete and fragmentary characteristics. However, most geographic information applications need geographic information data in valid time and space effectiveness strictly. So, although the geographic information data showing explosive growth, still unable to meet the demand of geographical information application. The rise of big data idea and technology brings a solution for this problem, it makes the role of geographic information data organization and management changed, from massive data storage and service extends to improve the utilization rate of the massive data. Under the background of Data Technology (DT) Era, it is possible to achieve spatial information continuous forecast and simulation by using intelligent data mining technology, and this approach will be the trend of the development of the geographic data technology. In this trend, put forward new requirements and challenges in the study of spatial information data organization model. Geographic spatiotemporal data intelligent organization technology aims for the information intelligent discovery. It emphasizes the description and expression of the internal relations and characteristics of geographic spatial data, which can enhance the ability of information discovery from geographic spatiotemporal big data, and solve the bottleneck of geographic spatiotemporal data in valid time and space for GIS applications.In this paper, in order to meet the needs of spatiotemporal raster big data organization, we presented an intelligent spatial information cube model, named Geo Raster Intelligent Cube (GRIC). GRIC can fit, smooth and fill the discrete and fragmentary spatiotemporal raster data, based on the potential data relationship of spatiotemporal raster big data, including timing relationship, spatial relationships, attributes semantic relationship and evolution relationship, which makes spatiotemporal raster big data satisfy more geospatial information applications. It briefly includes the following contents:(1) Modeling of GRIC modelUnder the background of DT Era, this page analysised human intelligence approach to information and knowledge and designed a data organization conceptual model named Intelligent Data Cube (IDC). Based on the IDC model and combined with the characteristics of spatiotemporal data, this page put forward GRIC model concept and feature definition, designed the struction of GRIC model, and expounded the basic principle of GRIC model.(2) Design and demonstration of key technologies and application methods based on GRIC modelThe design and verification of the key technologies and application methods based on GRIC model are based on the available technology. The key technologies and application methods are about multi-source spatiotemporal data integration organization, temporal, spatial and attribute semantic ontology description method, massive data storage and computing technologies, related information searching method, spatiotemporal information forecast and simulation mechanism, and so on.(3) Implementation and analysis of a prototype of remote sensing database based on GRICWith the support of the national project of "China high-resolution earth observation system", this paper has carried out the design and implementation of the GRIC prototype. The prototype system is based on massive multi-source satellite data and its derived information products as the test data. The functions of data storage, management, transmission, producing and simulation based on GRIC model are realized in the prototype system. Experiments for testing the efficiency of data transmission and calculation, and the effect of data simulation have been performed. The experimental results show that the GRIC model has obvious advantages in the transmission and calculation of the multi user concurrent cases and the spatiotemporal information forecast and simulation mechanism of GRIC model is feasible.