Abstract: Concerning the contradiction of the limited computer memory and the huge terrain data in terrain rendering, the tile-pyramid model was built up and the optimum size of the tile was analyzed. An efficient method of quadtree tile-index according to the relation of quardtree tiles and a thought of data compress based on "tile" were also put forward. The method not only analyzed but also optimized the interactive strategy between the massive terrain data and the three-dimensional display from the data pre-loaded, double-caching mechanism, multi-thread and memory. In combination with classic Geometry Clipmaps algorithm, the terrain rendering was accomplished. The experimental results show excellent rendering effect and good real-time performance, and prove the feasibility and efficiency of the optimized strategy.

Key words: massive terrain, tile-pyramid model, quadtree tile-index, beforehand loading, multi-thread

摘要： 针对海量地形绘制中有限的计算机内存与海量地形数据之间的矛盾，对海量地形数据采用瓦片金字塔模型进行了建模，分析了瓦片最优值的选取，根据四叉树瓦片拓扑关系提出一种高效的四叉树瓦片检索方式，并给出了基于瓦片的数据压缩思路与方法。对海量地形数据与三维显示的交互策略从数据预装载、双缓存机制、多线程技术、内存优化等方面进行了分析和优化，并结合经典的Geometry Clipmaps算法进行了地形绘制，实验结果具有较好的渲染效果和较高的实时性，证明了优化策略的可行性和有效性。

关键词: 海量地形, 瓦片金字塔模型, 四叉树瓦片索引, 预装载, 多线程