基于限制性四叉树LOD大规模地形预处理算法

Preprocessing LOD algorithm for large scale terrain based on restricted quadtrees

LOD(Level Of Detail,层次细节)技术是解决大规模地形实时渲染的关键技术之一,通过这种技术可以较好地简化场景的复杂度,减少图形显示的失真度,满足一定的实时性要求。传统的算法将四叉树和LOD技术相结合将大规模数字高程模型数据(DEM)进行分块,并对块内数据按照分辨率的大小分层存储。通过对四叉树的研究,在限制性四叉树的基础上引入预处理算法,提高了地形读取速度,增强了实时显示效果。该算法是基于限制性四叉树的一种高效的规则网格划分方法,内存开销少,降低了CPU的负担。实验结果表明该算法提高了地形导入的效率,能实现大规模地形的实时漫游。

LOD（Level Of Detail） which can reduce the complexity of scene preferably is the a key technology to solve the rendering problem of large scale terrain, which, LOD can alleviate the distortion of graphics and meet the request of real-time.Traditional algorithm,which combining quadtrees with Level Of Detail technology,partitions the DEM（Digital Elevation Model） of the terrain and stores the data of the block hierarchically by the resolution.In this paper,a preprocessing algorithm is introduced,which based on the restricted quadtrees.This method can increase the speed of loading terrain,and improve the effect of the real-time displaying, algorithm based on restricted quadtrees is an effective regular mesh refinement,which spends less memory and reduces the burden of CPU.Experimental results show that the algorithm improves the efficiency of terrain loading as well as supports real-time visualization of large-scale terrain rendering of online games well.