形状可控的烟雾动画

发布时间：2018-03-08 19:46

本文选题：烟雾动画　切入点：控制算法　出处：《浙江大学》2014年博士论文　论文类型：学位论文

【摘要】：烟雾控制一直是流体仿真领域的重要分支之一。可控的烟雾动画在影视特效和广告中的应用也越来越广泛。这种利用计算机仿真技术来制作动画特效的方式不仅解开了现实对艺术家或导演的束缚,让他们能够发挥天马行空的想象力,而且也为影视等娱乐产业节省了大量的拍摄成本,同时还能减少或者避免让演员暴露在某些危险场景中的可能性。因为这些重要的应用价值,烟雾控制也吸引了大量研究者的关注和投入。通过详细地分析国内外的研究现状,本文对烟雾控制的几个关键问题进行了深入的探讨,并提供了系统的解决方案。具体来讲,本文的贡献包括以下几个方面：·烟雾控制既包含路径控制又包含形状控制,为了能给动画师带来更多的便利,我们提出了一种统一的控制框架。该算法在同样的计算框架下既能实现路径控制,又能实现形状控制,还能实现两者结合的混合控制。为了实现这种统一的控制方式,我们把表示目标路径的空间曲线和表示目标形状的几何模型转化为对应的有向距离场,并基于这种有向距离场设计了三种控制力：路径控制力、边界控制力、以及基于中轴点云的形状控制力。路径控制力能让烟雾沿着目标的路径运动。边界控制力能将烟雾限制在指定的区域内运动。形状控制力则能驱使烟雾形成各种各样的目标形状。同时,为了提高形状控制的精度,我们还提出了一种自适应的散度调整策略,能够去除仿真中多余的烟雾,从而使得结果形状上的细节特征更加清晰。另外,我们还设计了一种混合涡旋粒子,能够根据所处的空间位置自适应地增强烟雾的湍流细节。实验也证明了我们的算法要优于有的形状控制算法和路径控制算法,以及通过简单修改将两者结合起来的混合算法。·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。因此,一个快速有效的计算框架会给动画师带来很大的便利,节省动画设计的时间。为此,我们提出了一种新颖的、针对可控烟雾动画的计算框架。在该框架下,我们采用基于离散正弦变换(DST)的泊松方程求解算法取代传统的PCG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,还能完成流体方程中对流项的求解。实验表明,我们的算法框架能得到与传统框架十分相似的结果,同时将计算速度提高了20倍以上。·针对形状和路径可控烟雾动画提出了一种过程式的湍流细节合成算法。为了在增强湍流细节的同时,避免对烟雾的控制效果造成不利的影响,我们设计了一种新颖的、随时间和空间变化的合成参数。该合成参数由涡旋速度和控制力共同计算得到,可以有效地控制过程式生成的湍流行为。使用合成参数后,仿真结果在得到丰富的湍流细节的同时,能很好的保持控制效果不受到大尺度的噪声带来的不利影响。通过和其它经典的过程式合成算法的比较,证明了该算法的有效性和可控性。
[Abstract]:Smoke control is one of the most important branch of fluid simulation field. The application of controllable smoke animation special effects in film and advertising is more and more widely. This kind of using computer simulation technology to make the animation way not only solved the reality of an artist or a director, so that they can play a powerful and unconstrained style imagination, but also saves the shooting a lot of the cost of the film and entertainment industry, but also can reduce or avoid the possibility of exposure actors in some dangerous scenes. Because of the important application value, smoke control has also attracted a lot of attention of researchers and investment. Through a detailed analysis of the status quo at home and abroad, this thesis deeply discusses several key the problem of smoke control, and provides the solution of the system. Specifically, the contribution of this paper is as follows: smoke control For both path control consists of shape control, in order to bring more convenience to the animators, we propose a unified control framework. This algorithm in the computational framework under the same path can achieve control, but also can realize shape control, hybrid control can realize the combination of the two. In order to realize the unified control mode we have said, the space curve and the geometric model of the target path that the shape of the object is transformed into the corresponding to the distance field, and based on the design of three kinds of control force to the distance field: path control, boundary control, and the axis of the shape of point cloud based on control path motion path control can. Let the smoke along the target. The boundary control can limit the movement of smoke in the designated area. The target shape control force can drive the smog of all kinds. At the same time, in order to improve the shape The shape control precision, we also propose an adaptive divergence adjustment strategy, can remove excess smoke simulation, which makes the results more clear on the details of the feature shape. In addition, we also design a hybrid vortex particle, according to the space position of the located adaptive enhanced turbulence. The details of the smoke we also prove the algorithm to algorithm and path control algorithm to control the shape is better than some, as well as through simple modification of hybrid algorithm combining the two. In the production of controlled smoke animation, often need to debug various parameters to get satisfactory results. Therefore, a fast and effective calculation framework will bring great convenience to the animators, save the animation design time. Therefore, we propose a novel computational framework, aiming at the smoke animation. In this framework, we use the Based on the discrete sine transform (DST) algorithm for solving the Poisson equation to replace the traditional PCG solver, and improved the existing vortex particle algorithm, the enhanced turbulence details at the same time, can solve the convection fluid equations. Experimental results show that the algorithm can get our framework is very similar with the traditional framework of the results, at the same time the calculation speed is increased by more than 20 times. The smoke controllable shape and a turbulent path animation details an algorithm for synthesizing process. In order to enhance the turbulent details at the same time, to avoid the smoke control effect to cause adverse effects, we design a novel synthesis parameters, with the change of time and space the synthesis parameters by the scroll speed and the control force is calculated, the turbulence behavior can effectively control the process of production. The use of synthetic parameters, the simulation results obtained in rich turbulence The flow details can also keep the control effect well without the adverse effects of large scale noise. Compared with other classic process synthesis algorithms, the effectiveness and controllability of the algorithm are proved.

【学位授予单位】：浙江大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TP391.41;TP391.9

【相似文献】