当前位置:主页 > 文艺论文 > 广告艺术论文 >

3D引擎中多层地形实时渲染及场景调度算法研究与实现

发布时间:2018-10-22 14:35
【摘要】: 随着计算机软硬件技术的不断发展,三维图形引擎也慢慢成为了计算机图形学研究的主要内容。三维图形引擎在虚拟现实、数字城市、影视广告、计算机辅助设计和科学计算可视化等各个方面起着十分重要的作用。通过三维图形引擎我们可以逼真地模拟人在自然环境中视觉、听觉及运动等。一个完整功能强大的三维图形引擎融合了计算机图形学、多媒体技术、人工智能、人机接口技术、数字图像处理、网络技术、传感器技术以及高度并行的实时计算技术等多个信息技术分支。 对此,本文从国内外的研究现状出发。对地形的几何建模和光学建模进行深入研究,并且在大规模场景的数据调度进行改进。从而给出大规模地形实时渲染的步骤和大规模场景数据调度的优化算法。 在大规模地形的渲染处理中,使用了基于组件式的多层地形算法。通过组件式的地形,我们可以将地形表现的更加的丰富。此外,组件式地形也是本文提出的多层地形渲染算法和LRU数据加载算法很重要的基础。组件式地形算法的效率将会直接影响到后面两个核心算法的效率。在组件式地形算法的基础上,通过多层地形算法,实现地形的多种精细程度,从而更加真实的模拟现实世界。同时避免了LOD算法消耗很高的问题,在保证真实性的情况下,减少各种消耗。 多层地形算法虽然降低了开销,但是却增加了内存消耗。简单说来就是通过空间来换取效率,为了减少空间消耗的影响,因此本文的第二个核心算法是LRU的数据加载。通过LRU算法可以尽最大可能的减少空间消耗的弊端。 在大规模场景的数据调度中,使用了基于LRU算法原理的数据调度算法。通过本文的算法,可以实现大规模场景的数据调度,同时还不需要担心由于内存大小对场景的限制,从而保证大规模场景中数据处理的流畅性。 此外,通过原创的三维图形引擎来实现本文的算法,从而证实算法的可行性。根据实验的结果说明算法能够很好的满足三维漫游的需求。 最后,在本文算法的基础上提出改进意见和进一步的展望。
[Abstract]:With the development of computer software and hardware technology, 3D graphics engine has become the main content of computer graphics research. 3D graphics engine plays a very important role in virtual reality, digital city, video advertising, computer aided design and visualization of scientific calculation and so on. Through the 3D graphics engine, we can simulate human vision, hearing and motion in the natural environment. A fully functional 3D graphics engine combines computer graphics, multimedia technology, artificial intelligence, man-machine interface technology, digital image processing, network technology, Sensor technology and high-parallel real-time computing technology and other information technology branches. In view of this, this article proceeds from the domestic and foreign research present situation. The geometric modeling and optical modeling of terrain are studied, and the data scheduling of large-scale scene is improved. The steps of real-time rendering of large-scale terrain and the optimization algorithm of data scheduling for large-scale scene are presented. In the rendering of large scale terrain, a multi-layer terrain algorithm based on component is used. Through the component of the terrain, we can be more rich terrain performance. In addition, the component terrain is also an important foundation of the multilayer terrain rendering algorithm and the LRU data loading algorithm proposed in this paper. The efficiency of component terrain algorithm will directly affect the efficiency of the latter two core algorithms. On the basis of the component terrain algorithm, the multi-layer terrain algorithm is used to realize the various degree of fineness of the terrain, so as to simulate the real world more truthfully. At the same time, the problem of high consumption of LOD algorithm is avoided, and all kinds of consumption are reduced under the condition of ensuring authenticity. Although the multilayer terrain algorithm reduces the overhead, it increases the memory consumption. In order to reduce the effect of space consumption, the second core algorithm of this paper is LRU data loading. The LRU algorithm can reduce the disadvantages of space consumption as much as possible. The data scheduling algorithm based on LRU algorithm is used in large scale scene data scheduling. The algorithm in this paper can realize the data scheduling of large-scale scene without worrying about the limitation of memory size on the scene so as to ensure the smoothness of data processing in large-scale scene. In addition, the algorithm is realized by the original 3D graphics engine, which proves the feasibility of the algorithm. The experimental results show that the algorithm can meet the needs of 3D roaming. Finally, on the basis of the algorithm proposed in this paper, suggestions and further prospects.
【学位授予单位】:电子科技大学
【学位级别】:硕士
【学位授予年份】:2009
【分类号】:TP391.41

【引证文献】

相关硕士学位论文 前3条

1 栗超;一种三维可视化系统的优化策略[D];燕山大学;2012年

2 梁雅港;火星探测任务三维仿真关键技术研究[D];解放军信息工程大学;2011年

3 钟鸣;三维场景的渲染技术研究与实现[D];武汉理工大学;2012年



本文编号:2287439

资料下载
论文发表

本文链接:https://www.wllwen.com/wenyilunwen/guanggaoshejilunwen/2287439.html


Copyright(c)文论论文网All Rights Reserved | 网站地图 |

版权申明:资料由用户8681f***提供,本站仅收录摘要或目录,作者需要删除请E-mail邮箱bigeng88@qq.com