CASIA OpenIR  > 毕业生  > 博士学位论文
应用GPU技术的建模、模拟和绘制的实例研究
其他题名Case Studies on Modeling, Simulation and Rendering Using GPU Programming Technique
王宁
学位类型工学博士
导师胡包钢
2011-05-27
学位授予单位中国科学院研究生院
学位授予地点中国科学院自动化研究所
学位专业计算机应用技术
关键词计算机建模 模拟和绘制 图形处理单元 Modeling Simulation And Rendering Graphics Processing Unit
摘要随着图形处理器(graphics processing unit,GPU)性能的不断提升和可编程灵活性的不断提高,越来越多的计算机图形学建模和模拟工作可以在GPU上有效的实现。图形学建模和模拟工作主要包括模型方法和模拟绘制两个方面的内容,本文的研究工作在这两个方面都进行了一定程度的创新:一是模型方法上,所有的建模方法和模型系统都是根据具体应用问题首次提出,具有一定的特色和新颖性。二是模拟和绘制工作均尽可能在GPU上进行,充分利用了GPU的高性能和可编程灵活性,在提高了性能的同时又减轻了CPU的计算负担,避免了数据在CPU和GPU之间的传输。本文具体的工作内容论述如下: 对于自然界中的风沙运动,提出了一种基于物理的建模和模拟方法。我们的建模工作主要借鉴了风沙物理学中单个沙粒的运动力学机制。该力学机制在风沙物理学中已经受到广泛的研究,而在计算机图形学中却是第一次得到应用。我们将最重要的沙粒运动机制引入模型当中 ,例如跃移、相继跃移和崩塌,同时简化模型中难以承受的内容,例如植被模型和风场模型。我们将模拟和绘制工作在GPU上实现,从而可以展现三维场景中大量沙粒的实时模拟和绘制。通过仿真实验证明了我们的建模和模拟方法可以实时的展示许多重要的沙波纹演化特性。最后我们还提出了一种利用模拟的沙床几何来合成沙漠场景的方法,以及演示如何利用模拟的运动沙粒来实时绘制具有真实感的风沙流。 对于铅笔素描绘画,提出了一种基于笔触的、自动和交互式结合的铅笔素描建模和模拟系统。我们的模型和系统在以往一些工作的基础上,加入了人类的绘画行为习惯和绘画技法。我们还提出了一种新的采样方法来自动生成影线笔触。和前人的相关工作相比,该采样方法简单方便,且提高了采样点利用率和绘制效率。我们还将系统中相关的许多模拟和绘制工作在GPU上实现,在减轻了CPU负担的同时也使得未来大规模的笔触模拟和绘制成为可能。最后通过实验结果证明,该系统可以模拟出逼真的铅笔素描作品。即使无绘画基础的用户也可以在系统的辅助下,快速生成具有手绘效果的铅笔素描作品。由于系统是基于笔触的,相对于其它图像、体绘制、纹理映射等铅笔素描生成方法,具有符合实际绘画创作过程、易于编辑修改等优点,因而具有更高的实用价值。
其他摘要With the performance and programmable flexibility of graphics processing unit (GPU) improved continuously, a growing number of computer graphics modeling and simulation works can be implemented on it. The modeling and simulation works are mainly include two aspects: one is modeling method, another is simulation and rendering. This thesis carried out innovation in both aspects. All the modeling methods in the thesis are first proposed according to specific applications. We also make the simulation and rendering on GPU to utilize its high performance and programmable flexibility. Simulation on GPU can also reduce the burden of CPU and avoid data transmission between CPU and GPU. Details of the thesis are as follows: We present a physically based modeling and simulation method for aeolian sand movement. Our method is based on the mechanical behavior of single sand grains, which is widely studied in the Physics of blown sand, but is first applied in computer graphics. We took significant mechanisms of sand transportation into our model, such as saltation, successive saltation and collapsing, while simplified the vegetation model and wind field model to make the simulation feasible and affordable. To accelerate the simulation process, we implemented the proposed method on the programming graphics processing unit. Finally, we proved that our method can reflect many characteristics of sand ripple evolution through several demonstrations. We also gave several synthesized desert scenes made from the simulated height field to display its significance on application. We present a stroke based modeling and simulation system that can generate pencil drawings both automatically and interactively. On the basis of existing works, we add some human drawing behaviors and drawings skills to our system. We also proposed a new sampling method to automatically generate hatching strokes. Compared to previous related work, the proposed sampling method is simple, and can improve rendering efficiency and utilization of sampling points. We implemented the simulation and rendering mainly on GPU to reduce the burden of CPU. This also makes large number of strokes simulation and rendering become possible. Finally, we showed that our system can fast generate realistic and vivid pencil drawing pictures. With the aid of the system, even non-professional users can fast design handcrafted pencil drawings. Compared with other image based, volume based and texture mapping based pencil d...
馆藏号XWLW1669
其他标识符200618014629080
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/6357
专题毕业生_博士学位论文
推荐引用方式
GB/T 7714
王宁. 应用GPU技术的建模、模拟和绘制的实例研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2011.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
CASIA_20061801462908(5562KB) 暂不开放CC BY-NC-SA请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[王宁]的文章
百度学术
百度学术中相似的文章
[王宁]的文章
必应学术
必应学术中相似的文章
[王宁]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。