面向装配的飞行器超大模型实时可视化技术

摘要

针对飞行器大数据量CAD模型实时绘制困难的问题,提出一种LOD自动批处理生成以及实时自适应绘制方法.以零部件为处理对象以保留装配树信息,结合模型分割完成超大模型简化,并根据计算机存储能力、实时绘制能力以及CAD模型特点,实现了定量的LOD自动批处理生成；以精确遮罩查询为基础,根据计算机实时绘制负载和CAD模型面片密度,动态地调整LOD精度等级,使得各个模型的精度基本一致,避免了传统算法的弊端,实现了LOD实时自适应绘制,并进行了优化.实验结果表明；采用文中技术处理千万级三角面片的模型约1h完成LOD生成,并可在普通计算机上实现实时的自适应绘制.%In order to solve real-time rendering of massive aircraft CAD models, automatic batch LOD generation and real-time adaptive rendering methods are presented. By the methods, CAD parts are treated as handling targets to keep assembly tree, together with split algorithm to simplify massive model. Then model storage, rendering ability of computer and characteristics of CAD parts are taken into consideration for quantitative LOD generation. LOD levels are adjusted according to rendering load and mesh density based on accurate occlusion query, so the accuracies and errors among different parts are consistent, which also avoid drawbacks of traditional methods. Finally the methods are optimized for speed. The experimental results show that massive CAD model with tens of millions triangles can be finished LOD generation within one hour and rendered in real-time on personal computers.