首页 | 本学科首页   官方微博 | 高级检索  
     

一种用于实时体绘制系统的自适应采样算法
引用本文:夏放怀,沈振康,唐朝京,谌海新,Juergen Hesser,Reinhard Mnner.一种用于实时体绘制系统的自适应采样算法[J].电子学报,2002,30(3):367-371.
作者姓名:夏放怀  沈振康  唐朝京  谌海新  Juergen Hesser  Reinhard Mnner
作者单位:1. 国防科技大学电子工程学院,湖南长沙 410073;2. Lehrstuhl für Informatik V,Universitt Mannheim,B6,26,D-68131 Mannheim,Germany
摘    要:本文讨论了一种用于体绘制的、支持基于空间跳跃和自适应光线终止等优化算法的高效硬件结构及一种加速半透明物体绘制的新技术.这种硬件结构用于绘制规模为2563体元的数据时可达到70Hz的帧频,但这样的帧频只是在体元的透明度都被置为0(完全透明)或1(完全不透明)的情况下才能取得.引入新的加速技术后,绘制半透明物体的帧频也可接近上述数值.新的加速技术采用自适应采样步长减少体绘制过程中冗余的体元重采样,从而使算法的复杂度和对存储器的带宽要求大大减小.新的加速技术对半透明体数据可取得高达4.7倍的加速比.采用新的加速技术需要在开始交互绘制前对体数据进行距离编码.优化后的距离编码对有8兆体元的体数据的预处理时间仅需要8~30秒.

关 键 词:体绘制  可视化  算法  自适应  距离编码  
文章编号:0372-2112(2002)03-0367-05
收稿时间:2000-03-05

An Adaptive Sampling Based Algorithm for Real-Time Volume Rendering System
Juergen Hesser,Bernd Vettermann.An Adaptive Sampling Based Algorithm for Real-Time Volume Rendering System[J].Acta Electronica Sinica,2002,30(3):367-371.
Authors:Juergen Hesser  Bernd Vettermann
Affiliation:1. Institute of Electronic Engineering ,National Univ.Of Defense Tech.,Changsha,Hunan 410073,China;2. Lehrstuhl für Informatik V,Universitt Mannheim,B6,26,D-68131 Mannheim,Germany
Abstract:This paper presents a new volume rendering acceleration algorithm for semi transparent volume objects.The volume rendering pipeline processor VGE provides full support for the popular ray casting acceleration techniques such as adaptive ray termination and space leaping,thus it achieves frame rate up to 70Hz.Nevertheless,such high frame rates can only be achieved in the case that the volume elements are mapped either opaque or transparent.The new acceleration algorithm is designed to relax the dependence of the frame rate on the selected opacity mapping functions.The new algorithm uses an adaptive resampling strategy to replace the equidistant resampling,hence the time complexity is reduced and the memory bandwidth requirement is lowered.Experimental results show that the new algorithm can accelerate the rendering of semi transparent volume objects by up to 4 7 times.Like space leaping,the new algorithm needs to encode the adaptive sampling interval in a preprocessing stage.The preprocessing can be implemented in less than 30 seconds for volume data with 8 million voxels by heuristic algorithmic optimizations.
Keywords:direct volume rendering  visualization  algorithm  adaptive  distance coding
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《电子学报》浏览原始摘要信息
点击此处可从《电子学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号