大规模三维地形可视化算法研究进展  

Research Evolvement of Large Scale Terrain Visualization Algorithm

作　　者：张慧杰[1] 孙吉贵[1] 刘雪洁[1] 李泽海[1] 

机构地区：[1]吉林大学计算机科学与技术学院

出　　处：《计算机科学》2007年第3期10-16,共7页Computer Science

基　　金：国家自然科学基金(60473003);教育部"新世纪优秀人才支持计划";博士点基金(20050183065);东北师范大学自然科学青年基金(20051003);吉林省杰出青年基金(20030107)资助。

摘　　要：大规模地形可视化是大型户外环境模拟不可缺少的组成部分,也是近年来可视化领域的研究热点,在游戏、仿真、虚拟现实、地理信息系统等领域有着广泛的应用。本文重点讨论了国内外学者在该领域的研究方法和最新研究进展以及尚未解决的问题。从数据拟合和模型简化两个方面叙述了自适应地形可视化建模方法,根据对现代图形硬件是否友好,将地形模型简化算法归纳为面向CPU的细粒度LOD算法和面向GPU的粗粒度LOD算法两类,同时描述了建模过程中存在的空间不连续问题以及各种解决方案,详细阐述了支持大数据集绘制的out-of-core技术,最后总结并分析了地形可视化建模领域的发展趋势和今后的研究重点。In recent years, large-scale three-dimensional terrain visualization is indispensable to outdoor environment simulation and a surge of research in the area of visualization. It is widely applicable for games, simulators, visual realization, GIS and other areas. This paper discussed research methods of many scholars and new research evolvement and some issues that still exist in terrain visualization area. On both data fitting and model simplification sides, these methods of adaptive terrain visualization modeling are described. According to be friendly for modem graphics hardware or not, theseterrain model simplification algorithms are divided into CPU oriented fine granularity LOD and GPU oriented coarse granularity LOD algorithm, as well as, this paper described space-time discontinuity problems, their various settlements and out-of-core technology that supports large-scale datasets. Finally, it is summarized and analyzed that the development current and future research emphasis in the terrain visualization modeling area.

关 键 词：大规模地形可视化 多分辨率模型 面向GPU的LOD算法 粗粒度 Out-of-core技术 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]