一种改进的基于CUDA的纹理映射和光线投射结合的体绘制算法
针对传统的基于GPU的光线投射算法绘制效率较低的问题,利用CUDA架构的并行计算特性和对三维纹理的处理能力进行改进和优化。将体数据映射为三维纹理,利用CUDA三维数组进行存储与绑定,纹理拾取的浮点返回值利用线性滤波进行平滑。在传输函数的设计中引入中心差分梯度幅值增强对体数据边界面的绘制效果。每条光线的求交及颜色积累采用并行计算,按照由前向后进行颜色及不透明度累积。设置不透明度阈值,采用不透明度提前终止加速绘制。实验结果表明,绘制速度较传统的基于GPU算法有10%的速度提升,绘制效果也有很大的改善。...
Saved in:
| Published in | 计算机应用研究 Vol. 32; no. 6; pp. 1884 - 1887 |
|---|---|
| Main Author | |
| Format | Journal Article |
| Language | Chinese |
| Published |
常熟理工学院计算机科学与工程学院,江苏常熟,215500
2015
|
| Subjects | |
| Online Access | Get full text |
| ISSN | 1001-3695 |
| DOI | 10.3969/j.issn.1001-3695.2015.06.065 |
Cover
| Summary: | 针对传统的基于GPU的光线投射算法绘制效率较低的问题,利用CUDA架构的并行计算特性和对三维纹理的处理能力进行改进和优化。将体数据映射为三维纹理,利用CUDA三维数组进行存储与绑定,纹理拾取的浮点返回值利用线性滤波进行平滑。在传输函数的设计中引入中心差分梯度幅值增强对体数据边界面的绘制效果。每条光线的求交及颜色积累采用并行计算,按照由前向后进行颜色及不透明度累积。设置不透明度阈值,采用不透明度提前终止加速绘制。实验结果表明,绘制速度较传统的基于GPU算法有10%的速度提升,绘制效果也有很大的改善。 |
|---|---|
| Bibliography: | 51-1196/TP To the problem of low efficiency of ray casting algorithm based on the GPU, used CUDA architecture to improve- ment and optimization algorithm with its parallel computing and processing ability of 3D texture. It mapped the volume data to 3 D texture, stored and binded with 3 D a~ray, smoothed floating-point return value of volume fetching with linear filtering. It de- signed transfer function with central difference gradient amplitude to enhance the rendering quality of boundary surface. The in- tersection and RGB accumulation of each ray with parallel computing, RGB and opacity accumulation with the direction of front to back along the ray. Accelerating the render with the method of early opacity termination according to the opacity threshoid. The experimental results show that rendering speed than traditional algorithm based on GPU has about 10% speed increase, rendering effect also has greatly improved. Zhu Shi, Chang Jinyi （School of Computer Science ＆ Engineering, Changshu Institute of Technolog |
| ISSN: | 1001-3695 |
| DOI: | 10.3969/j.issn.1001-3695.2015.06.065 |