文章摘要
基于实时LOD简化绘制的渲染优化方法
Rendering optimization based on real-time LOD simplification
  
DOI:
中文关键词: 虚拟展示  层次细节模型  通用并行计算架构
英文关键词: virtual demonstration  levels of detail  compute unified device architecture
基金项目:
作者单位
刘晓,刘镇,梅向东 江苏科技大学 计算机科学与工程学院江苏 镇江 212003 
摘要点击次数: 949
全文下载次数: 
中文摘要:
      随着虚拟展示应用在移动网络中的不断发展,在面对大量移动智能终端的并发请求时,服务端的计算负载成为了实时渲染的主要瓶颈,针对如何优化服务端渲染计算提高渲染效率,提出一种基于场景实时LOD简化绘制的渲染优化方法.方法在服务端实时绘制过程中根据用户的交互状态,实时简化绘制三维模型的相应分辨率细节层次模型,控制渲染图像质量,避免不必要的图形绘制,降低渲染计算负载.三维模型简化由CUDA实现并行处理,通过空间八叉树进行划分并按照顶点重要程度进行模型并行简化,控制场景渲染过程中的几何复杂度.实验证明,方法能够有效降低渲染耗时,保证用户进行交互浏览时的实时性.在移动网络环境中,方法能够有效降低服务端负载压力,提高渲染效率.
英文摘要:
      With the virtual demonstration development in mobile network, the server side’s computing load becomed the major bottleneck for real-time rendering when facing too much mobile intelligent terminal requests at the same time. In order to resolve this issue ,a real-time LOD simplification method for rendering optimizing was comed. For this method, when server side was working on real time rendering it based on user interaction to render appropriate three-dimensional simplified models, controlled rendering image quantity to avoid unnecessary image rendering and reduced rendering computing load. The level of details models simplify was implemented by GPU parallel processing, through space octree to divide and based on vertex significance level to implement parallel simplification, controlled scene rendering geometry complexity. From experimental results ,this method reduce rendering consumption effectively and guarantee user interaction fluency. In mobile network this method reduce server side loading pressure and improve resource utilization.
查看全文   查看/发表评论  下载PDF阅读器
关闭