Study Of DEM's Management And Scheduling Key Technologies Based On Large-Scale Dataset

With the rapid development of science and technology, interactive visualization of very large scale grid digital elevation models have a wide range of applications in many fields, such as:virtual reality, battlefield simulation, robot navigation, space exploration and so on. interactive visualization of very large scale terrain data brings up a wealth of problems. The main problem in real-time graphics is rendering efficiency. To best exploit the rendering performance, the scene complexity must be reduced as m uch as possible without leading to an inferior visual representation. Therefore, the geometry simplification must be controlled by an approximation error threshold. Another way to increase efficiency is the use of different levels of detail (LODs) for different areas of the visible scene. Objects are displayed in lower resolutions-with higher approximation errors-the farther away they are from the view focus.Because LODs can't reduce the scale of original data, for solving the problems of rendering efficiency, we'll discuss some issues such as:terrain modeling, error approximations and model simplifying algorithms and so on.First of all, some backgrounds, research status and theorys were introduced in Chpter 1&2, and we discusseed the advantages and shortcomings of two similar approaches here—the Pyramid Model and the Multi-Resolution Model.Secondly, a new constructing method—Multi-Resolution Model based on Vector, VMRM—was presented in Chapter 3. Some definitions, theorems, characters and data structure of VMRM were given, and the reliability and rationality of VMRM were also proved in this section.Thirdly, by analyzing two cracks-fixing methods, a new cracks-fixing approache was presented in this paper, and its feasibility and reliability were verified by experiments.Fourthly, by discussing the advantages and disadvantages of two error approximations—Hierarchical Geographical Error and Divergence Function Error, we partly presented there error functions of VMRM and experiment results in this paper.Lastly, by researching the shortage of Divide and Conquer, we studied the Particle Swarm Optimization (PSO) and then applied it in the model simplifying. Through researching the definitions, appraisals and speed updating of particle, we tested PSO's convergence and reliability in our reserches.