Research On Construction And Optimization Of Large Scale 3D Modeling For Realtime Rendering

As the development in computer technique especially in computer-graphics technology, the large scale real scene digital simulation is more and more widely used. How to quickly construct real-time rendering of the real large scale real scene model have become an important research topic. Just taking feasibility into consideration, as long as the computer hardware have enough computing power, enough image processing ability, it can show the real effect of enough detail. For the average animation character, use enough powerful rendering engines, given plenty of time so that you can gain final reality flowery rendering. But in real-time application requirements the system's each frame image and each frame processing results must be real-time rendering out. Generally speaking, for one event triggered, it must offer response and real-time rendering out the interactive result. So on the basis of certain hardware, how to make it through that keeping the large scale real scene realistic as well as real-time, has become a very important research goal in virtual reality field.This paper based on Central China Normal University digital campus model construction, combined 3Ds Max and Multigen Creator with their respective advantages for the model construction.First, modeling steps and classification. This part will lay the foundations for the following data optimization, it is also the premise of optimizing the data quantity and improves the visual effect at the same time,. Overall planning for the model to determine the model refinement level, and then according to the different requirements of refinement in different software for construction, Refining degrees demanding in 3Ds Max in construction, refining the low degree of construction in Creator. In the process, we should also use some strategies, and also need to make improvement in necessary modeling method, by adopting a certain optimization disposal strategy.Second, for each module construction after the model is partitioned into small blocks, when constructing the corresponding modeling method can be used in accelerating modeling speed while improving modeling effect. After module construction, it is time to put model conversion and integration. Then we need to transform 3Ds Max format into Openflight format supported by Creator.Some problems maybe appear during conversion process such as model disorder, texture flicker, data loss and so on, which will lead to the model cannot be loading by rendering system, or visual effect disorder.Third, Model design optimization, after the model incorporates together, it can be optimized from the following several aspects. First of all, is the amount of data model optimization? This aspect use the cleanup of redundant, integration the surface using adopting texture technology, instantiation technology, multiple level of detail model, external references technology, etc. And then anew in accordance with regional organizations node structure, accelerate rendering speed and efficiency. The last is visual effect optimization, it is mainly eliminate the process of confusion appeared in integration, and dark phenomenon.This paper makes a comparison and analysis the data quantity before and after optimum scene, usage of resources when roaming, visual effect and other factors, it turned out that this paper adopted the large scale real scene model construction and optimization method is feasible, and it has achieved favorable results.