一种基于松散八叉树的复杂场景可见性裁剪算法  

Visibility Culling for Complex Scene Based on a Loose Octree Structure

作　　者：梁爱民[1] 梁晓辉[1] 于卓[1] 

机构地区：[1]北京航空航天大学计算机学院虚拟现实技术国家重点实验室,北京100083

出　　处：《计算机辅助设计与图形学学报》2007年第12期1593-1598,共6页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家"八六三"高技术研究发展计划(2006AA01Z333);国家自然科学基金(60533070)

摘　　要：针对传统八叉树方法的不足,在采用松散八叉树组织场景、利用八叉树空间划分优点的同时弥补其局限性.为提高遮挡查询效率,将子节点依视点排序,针对复杂场景采用双层裁剪技术以进一步提高性能.实验结果表明,文中算法对深度复杂度高、面片数量大的复杂场景具有较好的裁剪效率,能够很好地满足实时绘制的要求.To eliminate the drawbacks of traditional octree structure and make full use of its advantages in visibility culling, we organize the scene into a loose octree structure. The nodes of the hierarchy are processed into a front-to-back order in order to improve performance. By the analysis to the complex scene, a double-layer visibility culling algorithm is proposed, and the algorithm combines view frustum culling, detail culling and occlusion culling methods. The results of experiments prove its high efficiency for handling complex scenes with large amount of triangles in high depth.

关 键 词：遮挡查询 可见性裁剪 松散八叉树 实时绘制 复杂场景 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]