A Research On Pyramid Model Construction And Data Processing Technology Of Terrain Data In Virtual Terrain Environment

As we all know, a global-based multi-resolution terrain visualization system depends greatly on the efficiency of data access, especially massive data. In view of this, we put emphasis on some problems related with massive data, such as how to construct a data pyramid model, how to process massive terrain data as fast as possible, how to optimize the processing algorithm, and how to design a parallel policy for data processing.The following is what we have done in this thesis:1.Design the global three-dimensional terrain data organization model which is based on the structure of Pyramid .Choose the best one among Pyramid model's layer and block division methods. Design a common layer and block division schema suiting many of the three-dimensional terrain data's display algorithms and hierarchical scheduling mechanisms.2.In order to satisfy the storage requirements of three-dimensional terrain data, respectively construct the three-dimensional terrain data's theory and physical storage models base on files and database types. Solve the efficient management problem of the physical table in the storage mode of database by using oracle partition technology.3.Design the workflow of three-dimensional data processing. Make a lot of research and experiments on several key issues such as data block reading, data re-sampling and texture compression. Implement the efficient operation of oracle database through OCCI. And improve the speed of data migration with the data pump technology of oracle.4.Make the designs and experiments on the key technology of using multi-threading tech, the efficient management of memory and the simplicity of image data complementarity. And improve the three-dimensional terrain data processing efficiency.5.Carry out the exploratory study on the method of multi-computer parallel processing; propose the task division method which reduces the conflicts in the data paralleling processing. Design the flow of the multi-computer data parallel processing.6.Verify the method designed by this article through program designations and experiments.