高级检索
杨贲, 金小刚, 王明. 采用离散正弦变换的快速可控的烟雾动画[J]. 计算机辅助设计与图形学学报, 2015, 27(7): 1230-1237.
引用本文: 杨贲, 金小刚, 王明. 采用离散正弦变换的快速可控的烟雾动画[J]. 计算机辅助设计与图形学学报, 2015, 27(7): 1230-1237.
Yang Ben, Jin Xiaogang, Wang Ming. Fast and Controllable Smoke Animation Based on DST[J]. Journal of Computer-Aided Design & Computer Graphics, 2015, 27(7): 1230-1237.
Citation: Yang Ben, Jin Xiaogang, Wang Ming. Fast and Controllable Smoke Animation Based on DST[J]. Journal of Computer-Aided Design & Computer Graphics, 2015, 27(7): 1230-1237.

采用离散正弦变换的快速可控的烟雾动画

Fast and Controllable Smoke Animation Based on DST

  • 摘要: 在制作可控的烟雾动画时常常需要反复调试各种参数.以得到满意的效果.为此,提出一种针对可控烟雾动画的计算框架,以节省动画师设计时间.该框架采用基于离散正弦变换的泊松方程求解算法取代传统的预处理共轭梯度求解,并改进了现有的涡旋粒子算法,使其在增强湍流细节的同时完成流体方程中对流项的求解.实验结果表明,文中的框架能得到与传统框架十分相似的结果,但计算速度可提高20倍以上.

     

    Abstract: When creating controllable smoke animation, animators often need to adjust various parameters to achieve a satisfactory effect. To this end, we propose a novel computing framework to save animators’ time and effort in designing controllable smoke animations. Under this computing framework, a Poisson equation solution algorithm based on discrete sine transformation has been employed to replace the traditional preconditioned conjugate gradient algorithm. In addition, an improved vortex particle algorithm has been introduced to get the solution of convective teams in fluid equations while enhancing turbulence details. Experiments show that the presented framework is over 20 times faster than the traditional framework while producing similar results.

     

/

返回文章
返回