基于四叉树的缓存复用机制地形算法-《计算机技术与发展》

文章信息/Info

Author(s):: SONG Xing-shen[1]; QUAN Ji-cheng[1]; ZHAO Xiu-ying[1]; WANG Hong-wei[1]; WANG Yu[2]

摘要:: 为了减少地形数据的冗余，实现缓存复用，在研究了Geo-CLipmap算法的基础上，针对其采用嵌套网格一次只能渲染单张高程图的方案，提出了一种块状四叉树的数据结构，同时分开存储节点的位置信息和高程信息，有效地减少了数据大小和数据交换，实现了顶点和索引缓存的复用；利用掩膜填充不同分辨率网格之间的缝隙，避免了加入光照后的几何失真，并使用几何过渡消除几何体的跳变。经过实验测试，该算法能有效地压缩数据量，在稳定帧速的情况下，实现地形的无缝渲染。

Abstract:: In order to decrease the redundancy in terrain data and reuse the cache,based on the research of Geo-Clipmap algorithm,which only renders single height map by using nested meshes,a data structure of chunk quadtree is proposed,which stores the position informa-tion and evaluation information of nodes separately,efficiently subtracting the size of datum and the transmittal between them by reusing vertex cache and index cache for rendering. By using masks filled in the seams between multiple resolution meshes,the algorithm success-fully avoids geometry aliasing under illumination,meanwhile,it eliminates popping by using geometry morphs. As the experiment shows, this algorithm can effectively compress the size of data and render terrain seamlessly while maintaining the frame rate smoothly.