基于GPU编程的地形可视化

A Terrain Visualization Algorithm Based on GPU Programming

由于地形模型固有的复杂性,致使计算机硬件水平一直难以满足大规模地形模型的实时显示需求。为了在现有的硬件水平上实现地形模型的快速绘制,在对传统的ROAM算法进行改进的基础上,提出一种基于GPU编程的地形可视化算法,实现了视点依赖的大规模地形的快速可视化。该算法首先基于改进的ROAM(real-time optimallyadaptive meshes)算法生成视点依赖的优化连续LOD模型;然后用GPU编程计算顶点的变换、法向量、纹理坐标、纹理采样和面元光照;最后完成地形的着色。实验结果表明,利用GPU编程不仅能有效提高算法速度,而且能实现较大规模地形的实时漫游。

Due to the inherent complexity of terrain models, the development speed of computer hardware is far from satisfaction for real-time rendering of large scale terrain model. To resolve this problem, we enhance the traditional ROAM algorithm and therefore realize view-dependent real-time rendering of large scale terrain model with GPU programming. Here, the enhanced ROAM （real-time optimally adaptive meshes） algorithm is employed to create view-dependent continuous LOD models, and the GPU programming is employed to calculate the vertices＇ transform, normal vector, texture coordinate, texture sampling and fragment lighting and accomplish the rendering for the terrain in the end. Experimental results show that, GPU programming is an effective way to improve the algorithm performance and to achieve real-time roaming on relatively large scale terrain.