北京大学学报(自然科学版)

一种基于数据分块的海量地形快速漫游方法

马照亭, 潘懋, 胡金星, 吴焕萍, 王占刚   

  1. 北京大学地球与空间科学学院,北京,100871
  • 收稿日期:2003-05-29 出版日期:2004-07-20 发布日期:2004-07-20

A Fast Walkthrough Method for Massive Terrain Based on Data Block Partition

MA Zhaoting, PAN Mao, HU Jinxing, WU Huanping, WANG Zhangang   

  1. School of Earth and Space Sciences, Peking University, Beijing, 100871
  • Received:2003-05-29 Online:2004-07-20 Published:2004-07-20

摘要: 大范围、海量地形数据的快速漫游是GIS、虚拟现实、游戏、仿真等领域的关键技术与难点之一。在目前地形数据简化算法的基础上,结合计算机硬件技术的发展,采用基于数据分块、部分数据常驻内存以及与视点相关的静态LOD模型实现了对任意格网大小海量地形数据的快速漫游。基于块节点的视景体裁剪、三角形条带组织以及边界裂缝消除、视觉光滑处理等手段的运用,提高了地形漫游的效率与可视化效果。

关键词: 海量数据, 地形漫游, 数据分块, 视景体裁剪, 静态层次细节

Abstract: Walkthrough in massive terrain data is a key technology and difficult problem of GIS, VR, game and simulation, etc. Based on current LOD algorithms and GPU technology of the computer, a fast walkthrough method with data block partition, out-of-core and static LODs is implemented. With view frustum culling based on nodes of block, triangle strip optimizing, cracks elimination and geomorphing, the efficiency and effect of the method are greatly increased.

Key words: massive data, terrain walkthrough, data block partition, view frustum culling, static level of detail

中图分类号: