基于CUDA的三维数据并行可视化

详细信息

CUDA-based Parallel Visualization of 3D Data

Department of Computer Science and Technology, Nanjing University of Science and Technology, Nanjing 210094, China

摘要: 科学计算可视化是发达国家20世纪80年代后期提出并发展起来的一个新的研究领域,它运用计算机图形学和图像处理技术,将三维数据转换为图形及图像在屏幕上显示出来并进行交互处理,主要包括面绘制和体绘制两种方法。光线投射算法是最经典的三维数据体绘制方法,图像质量较高,但计算时间较长,基于CUDA的编程技术利用GPU的多核并行运算功能可显著提高计算速度,实现大规模三维数据的实时可视化。
- 可视化 /
- 光线投射算法 /
- GPU /
- CUDA
Abstract: Visualization in scientific computing is a new field of study proposed and developed in developed countries in the 20th century and the late 80's. Using Computer Graphics and Image processing techniques,3D data can be converted to graphics or images displayed on screen for interactive processing. The main methods include surface visualization and volume rendering. Ray-casting algorithm is the most classical method of volume rendering methods about 3D data,which has higher image quality but longer calculation time. CUDA-based programming technique using the GPU hypercore parallel computing function,can significantly improve the calculation speed,and realize the real-time visualization of large-scale 3D data.
- visualization /
- ray-casting /
- GPU /
- CUDA