基于满二叉树分块策略的大规模数据场纹理映射体绘制算法

孙安玉,江贵平

doi:10.3969/j.issn.1007130X.2011.

计算机工程与科学 >

2011 , Vol. 33 >Issue 3: 57 - 61

DOI: https://doi.org/10.3969/j.issn.1007130X.2011.

论文

基于满二叉树分块策略的大规模数据场纹理映射体绘制算法

展开

(南方医科大学可视信息实验室,广东广州 510515)

孙安玉(1983),男,山东临清人，硕士生，研究方向为快速体绘制技术、医学可视化与增强现实。江贵平(1971),男，江西上饶人，博士，教授，研究方向为可视化与增强现实、微光机电系统应用。

收稿日期: 2010-09-18

修回日期: 2010-12-25

网络出版日期: 2011-03-25

基金资助

国家自然科学基金资助项目（30970866）

收起

Texture_Based Volume Rendering Using Full Binary Tree Partitioning for Large Datasets

Expand

(Visual Information Laboratory,Southern Medical University,Guangzhou 510515,China)

Received date: 2010-09-18

Revised date: 2010-12-25

Online published: 2011-03-25

Fold

摘要

针对纹理映射体绘制物理内存空间的限制，本文提出一种可在通用图形硬件上完成大规模数据场实时

体绘制的有效方法。该方法基于满二叉树纹理分块策略，利用GPU着色器可编程性，将纹理数据制作为一个

一维传递函数查找表和一个规模等同于体数据场的动态纹理工作集，有效提高了大规模数据场体绘制的实时

性。动态纹理工作集使用抽象分块与继承关系管理边界，通过计算层次化的方式降低了分块绘制时的计算及

动态纹理控制复杂度。实验表明，运用该算法可在普通PC上对远超过纹理内存的大规模体数据完成具有较好

实时性和较高质量的体绘制。

关键词： 满二叉树; 纹理分块; 体绘制; 三维纹理映射; GPGPU

本文引用格式

孙安玉,江贵平 . 基于满二叉树分块策略的大规模数据场纹理映射体绘制算法[J]. 计算机工程与科学, 2011 , 33(3) : 57 -61 . DOI: 10.3969/j.issn.1007130X.2011.

Abstract

In order to break through the limitation of GPU memory for the traditional

texture_based volume rendering, an effective technique utilizing 3D texture partitioning based

on full binary tree is presented for rendering largescale volume datasets interactively on

general purpose GPU . Using the Fragment Programming capability of commodity graphics cards,

the texture data is transferred to a 1D color lookup table and a working set of dynamic

texture datasets equivalent to the volume dataset size. The working set manages the boundaries

between blocks with the help of abstract partitioning and inherited attributes.The

experimental results show that this method is efficient for rendering largescale volume

datasets at an interactive rate on general purpose PCs．

Key words： full binary tree;texture partitioning;volume rendering;3D texture mapping;GPGPU

参考文献

［1］Kniss J,McCormick P,McPherson A,et al. Interactive TextureBased Volume Rendering for

Large Data Sets［J］.IEEE Computer Graphics and Applications,2001,21（4）:5261.
［2］Lee TH, Kim Y J, Chang J. High Quality Volume Rendering for Large Medical Datasets

Using GPUs［J］. Lecture Notes in Computer Science, Systems Modeling and Simulation: Theory

and Applications,2005,3398:663674.
［3］Agrawal A, Kohout J, Clapworthy G J,et al. Enabling the Interactive Display of Large

Medical Volume Datasets by Multiresolution Bricking［J］. The Journal of Supercomputing, 2010,

51(1):319.
［4］邹华. 基于可编程GPU的体绘制关键技术研究:［博士学位论文］［D］. 西安:西安电子科技大

学,2009.
［5］Zheng J, Ji H, Yang W. Interactive PC TextureBased Volume Rendering for Large Datasets

［C］∥Proc of the 1st Int’l Conf on Innovative Computing, Information and Control  Volume

II, 2006, 2:350353.

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献