A Research Of Optimizing Techniques For The 3D Scene Model Based On Creator

The one of aims of Virtual Reality technology is that the user could interact with the virtual world in a natural way as much as possible. The objects in Virtual Reality system give user the felling that they are real. In the current level of computer graphics, it's not a difficult to generate accurate and detailed computer images so far as there is sufficient computing time. However, if you want VR systems are real-time display, there are not only the effect of the scene should be considered, but also the speed of system response. Therefore, in order to ensure the high-quality graphics in VR display real-time, how to reduce the complexity of the graphics screen in the way of software, at the same time guarantee the accuracy of 3D scenes have been one of the main objectives of concreting a VR scene.The significance of optimizing techniques for the 3D scene model based on Creator is to provide an effective and efficient way to solve the contradiction between the quality and the speed of graphical display.This paper discusses methods of creation and optimization based on the software which is used to create 3D model such as MAYA and Creator taking the Huazhong Normal University Scene Simulation Management System and the Digital Campus 3D scene model for example. The main content is as follows:Firstly, optimizing techniques for the 3D scene model are classified into three types according to the basic information in OpenFlight model database. Remove and combine belong to optimizing techniques based on the characteristics of the model geometry. Optimizing techniques based on model attributes are focus on the optimizing of texture. Appropriate size and format as well as some texture techniques such as billboard help debase the EMS memory. Besides, appropriately organizing nodes, and LOD, Instance, External nodes are used to effective optimize the hierarchy structure of the database.Moreover, this paper sums up six strategies on how to optimize model data library referencing the results of previous studies and the experience gained during the work.Furthermore, this paper puts out some principles about how to product Discrete LOD model data referencing the results of previous studies and the experience gained during the work.Finally, this paper optimizes the model data according to the classification taking the Huazhong Normal University Digital Campus 3D scene model for example, then compares with the original system in the document, the load time, the EMS memory and so on. A conclusion is draw that these optimizing techniques adopted are feasible and necessary for real-time systems.