摘要
针对传统光线投射算法绘制速度慢和图形处理器(graphics processing unit,GPU)不能有效进行并行计算的缺点,文章提出一种基于包围跳跃的计算统一设备架构(compute unified device architecture,CUDA)光线投射算法。首先介绍了CUDA的编程模型和线程结构,然后用包围盒技术隔离体数据周围无效的空体素,减少投射光线的数目;利用光线跳跃技术,在包围盒内进行快速光线的合成,跳过透明的体素,减少大量体素的重采样;最后使用CUDA强大的并行处理计算的功能实现光线投射算法。实验结果表明,在保证图像质量的同时,绘制速度上比基于GPU加速的光线投射算法有14倍的提高,能够接近实时绘制,有很好的应用价值。
In view of the traditional ray casting algorithm for rendering speed slow and GPU （ Graphics Process- ing Unit） is not effective for parallel computing of the shortcomings. A CUDA （ Compute Unified Device Architec- ture） ray casting algorithm based on bounding leaping is proposed. Firstly the programming model and thread struc- ture of CUDA are introduced, and then useso the bounding box technique around the body data to separator the in- valid empty voxels, reducing the number of casting ray. Using the ray leaping technique, fast ray synthetic in the bounding box, skip transparent voxels, the number of voxels resampling is reduced. Finally the CUDA powerful parallel processing computing function is used to implement ray casting algorithm. The experimental results show that, the method is to ensure the image quality at the same time, the speed of rendering is 14 times increase than the ray casting algorithm based on GPU acceleration, near real-time rendering, has the good application value.
出处
《科学技术与工程》
北大核心
2014年第12期245-249,共5页
Science Technology and Engineering
基金
国家重大科学仪器设备开发专项(2012YQ05025004)资助
关键词
光线投射
体绘制
光线跳跃
包围盒
统一计算设备结构
ray casting volume rendering ray leaping bounding box compute untied device architecture （CUDA）
