The Research Of Raster Compression Coding Changes Based On Spatiotemporal Sequence

Raster data is a important form of GIS expression, it is frequently used in the fields ofland use and desertification, forestry virus control,fire spread,and so on. And with the cost ofremote sensing data acquisition is lower and lower,The research of distinction onmulti-temporal raster data is more and more important. However the amount of raster data isvery large, the spatio-temporal data also contain more than one data and there are a lot ofhistorical data. The purpose of this research is extracting differences of raster data,and savethe storage space of raster data,then to realize analysis and implementation of data griddifference history and space.The paper takes the three land use remote sensing image data of DaGang1988-2005asan example, based on raster data analysis method，studies the differences of spatio-temporalon land use. designs the multiple ground state is suitable for land use differential filecorrection model.The main contents of research:(1) Towards the requirement of the T_GIS management of raster data based on time line,analyzed the foundation of spatio-temporal raster data modeling and several common space ofraster data model, mentioned the design idea and the concept model of multiple ground stateand differential file correction model;(2) Spatio-temporal raster database has been designed from the conceptual and logical.Then we ameliorated the fast dynamic linear quad tree coding algorithm, The improvedalgorithm is applied to the two-dimensional run-length dynamic encoding, which is adoptedin multiple ground state and differential file correction model with amendments.(3) Then we have analyzed and researched on how to identify the differences betweenthe phases of raster data. The analysis and cumulative difference algorithms are improved.(4) In order to realize the grid data of space and time history back, is improved based ontwo-dimensional schedule difference accumulation dynamic compression encoding algorithm,namely the reduction of the ground state data.(5) In order to prove the presented algorithm is correct, design a analysis system ofspatio-temporal analysis differences of raster data on land use.