本文为荷兰乌得勒支大学（作者：Jack van der Drift）的硕士论文，共46页。

随着气候变化对干旱、强降雨和极地冰盖融化的影响越来越明显，研究更有效的流体模拟算法变得越来越重要。对于强降雨和极地冰盖融化等情况，能够使用现实生活中的激光扫描来预测景观会造成什么样的破坏以及我们需要在哪里采取预防措施将是很方便的。

本文的目的是提出一种在点云上直接进行流体模拟的方法。我们这样做是为了跳过通常所做的昂贵的重建步骤，简化标记粒子和目标之间的碰撞检测。我们的算法结合了MLS算法和欧拉流体模拟。对于每个网格点，我们计算MLS值，并使用这些值来确定网格单元的实体，防止标记粒子进入点云。实验表明，该方法可以直接在点云上进行流体模拟，但算法的性能与场景类型和使用的参数有很大关系。

With the effects of climate change becomingmore and more apparent in the form of drought, heavy rainfall and the meltingof the polar ice caps research into more efficient algorithms for fluidsimulation become more important. For cases such as heavy rainfall and themelting of the polar ice caps it would be convenient to be able to use reallife laser scans of landscapes to predict what damage they will do and where weneed to take precautions. The purpose of this thesis is to come up with amethod to directly use fluid simulation on point clouds. We do this to skip theexpensive reconstruction step which is normally done and simplify the collisiondetection between the marker particles and object. Our algorithm combines theMLS algorithm with a Eulerian fluid simulation. For each grid point we computethe MLS values and use these to determine how solid a grid cell is and keep themarker particles from going into the point cloud. Our experiments show that itis possible to directly perform fluid simulation on point clouds using ourmethod, but that the performance of our algorithm is highly dependent on thetype of scene and the used parameters.

1．  引言

2. 相关工作

3. 项目背景

4. 研究方法

5. 实验与结果

6. 讨论

更多精彩文章请关注公众号：