一种视点相关球面地形LOD算法

Algorithm for View-dependent Global Terrain LOD

在地形可视化领域,基于海量数据的全球地形3维可视化已经得到广泛关注和研究。在经典ROAM算法的基础上,提出了一种视点相关全球地形LOD模型SROAM。该算法采用正方体作为网格细分的顶层,正方体上每一个面分裂成两个三角形作为ROAM算法二叉树的根节点。给出了高程数据球面映射的方法,在动态误差计算过程中,考虑了地平线约束条件,进一步对网格进行简化。实验结果表明,此算法能实时动态地生成全球地形LOD模型,实现全球地形场景的平滑绘制。

Terrain visualization for whole global is a difficult problem for applications. We presented an algorithm for constructing view-dependent triangle meshes for whole global based on ROAM （ Real-Time Optimally Adapting Meshes）. The method, dubbed Spherical ROAM （ SROAM） , starts with a solid 3D cube that consists of only squares. Each square in this shape will be treated as a top-level square of ROAM algorithm. In this paper, the method of mapping terrain elevation data onto a global was given and the view-dependent dynamic node error calculation taking account of horizon of global was proposed. The experimental results showed that the algorithm proposed could dynamically generate view-depend multi-resolution global terrain in real time, and real time terrain rendering was realized.