一种基于GPU加速的三维分形实时渲染方法

发布时间：2018-04-26 22:33

本文选题：三维分形 + GPU　；参考：《激光杂志》2016年11期

【摘要】：绘制高质量三维分形图所需的计算量往往很大,难以达到实时的要求。研究了一种绘制三维分形的通用方法,即基于光线跟踪的距离估计算法。该算法非常适合在GPU上实现,利用GPU强大的并行处理能力,使三维分形的绘制速度显著提高。大量实验结果表明,在屏幕分辨率为1280x720时绝大部分三维分形图都已达到了实时绘制的要求。GPU与CPU的性能对比表明,即使在CPU最快的情况下耗时也是GPU的158倍,由此可见应用GPU的巨大优势。还验证了距离估计算法的通用性,对于不同类型的分形,如逃逸时间算法的分形、迭代函数系统的分形以及分形地形都可以适用。
[Abstract]:It is difficult to meet the requirement of real time because of the large amount of computation required to draw high quality 3D fractal images. In this paper, a general method of 3D fractal rendering, namely ray-tracing based distance estimation algorithm, is studied. This algorithm is very suitable to be implemented on GPU, and makes use of the powerful parallel processing ability of GPU to improve the rendering speed of 3D fractal. A large number of experimental results show that most of the 3D fractal images have met the requirements of real-time rendering when the screen resolution is 1280x720. The comparison of the performance of CPU and CPU shows that even in the fastest case of CPU, the time consuming is 158 times that of GPU. This shows the huge advantage of applying GPU. The generality of the distance estimation algorithm is also verified, which is applicable to different types of fractal, such as the fractal of the escape time algorithm, the fractal of the iterative function system and the fractal terrain.
【作者单位】：华南师范大学物理与电信工程学院;
【基金】：2014年广东省科技厅公益研究与能力建设专项资金(2014A040401076)
【分类号】：TP391.41

【相似文献】