计算物理 ›› 2005, Vol. 22 ›› Issue (4): 365-370.

• 研究简报 • 上一篇    下一篇

由扩散张量导出的各向异性扩散模型的隐式数值模拟

刘朝霞1, 常谦顺2   

  1. 1. 北京应用物理与计算数学研究所, 北京 100088;
    2. 中科院教学与系统科学研究院应用数学所, 北京 100080
  • 收稿日期:2004-03-17 修回日期:2004-08-30 出版日期:2005-07-25 发布日期:2005-07-25
  • 作者简介:刘朝霞(1974-),女,山东,讲师,从事数值分析与计算物理研究,北京8009信箱28分箱100088.

Implicit Numerical Simulations of an Anisotropic Diffusion Model Driven by Diffusion Tensors

LIU Zhao-xia1, CHANG Qian-shun2   

  1. 1. Institute of Applied Physics and Computational Mathematics, Beijing 100088, China;
    2. Academy of Mathematics and System Sciences, Chinese Academy of Science, Beijing 100080, China
  • Received:2004-03-17 Revised:2004-08-30 Online:2005-07-25 Published:2005-07-25

摘要: 研究了由扩散张量导出的各向异性扩散的图像处理模型,并构造了隐式差分格式,形成了有13条对角线的大型稀疏矩阵.利用代数多重网格法求解了这个线性代数方程组.并进行了数值试验.

关键词: 图像处理, 扩散张量, 各向异性扩散, 代数多重网格法

Abstract: A model of image processing for anisotropic diffusion driven by diffusion tensors is investigated. An implicit difference scheme is constructed. A large sparse matrix with 13 diagonals is formed with which general iterative methods don't work well. With algebraic multi-grid method (AMG), linear system of equations are solved.Numerical experiments are shown.

Key words: image processing, diffusion tensor, anisotropic diffusion, algebraic multi-grid method (AMG)

中图分类号: