J4

• 数学 • 上一篇    下一篇

非局部偏微分方程去噪模型

芦碧波, 郭志昌, 曹 杨, 李美慧, 胡歆瑶   

  1. 吉林大学 数学研究所, 长春 130012
  • 收稿日期:2007-10-12 修回日期:1900-01-01 出版日期:2007-11-26 发布日期:2007-11-26
  • 通讯作者: 芦碧波

Nonlocal Partial Differential Equation Model for Image Denoising

LU Bibo, GUO Zhichang, CAO Yang, LI Meihui, HU Xinyao   

  1. Institute of Mathematics, Jilin University, Changchun 130012, China
  • Received:2007-10-12 Revised:1900-01-01 Online:2007-11-26 Published:2007-11-26
  • Contact: LU Bibo

摘要: 针对传统扩散模型中的边界模糊问题, 提出一种基于非局部信息的非线性偏微分方程去噪模型. 在新模型中引入非局部梯度以探测图像边界, 在区域内部和边界处以不同的速度进行扩散, 从而达到去噪和边界保持的效果. 数值计算采用单边格式计算非局部梯度, 以防止离散情形下边界信息的损失. 理论分析和数值实验结果表明了此模型的有效性.

关键词: 图像去噪, 非局部信息, 偏微分方程

Abstract: A nonlinear partial differential equation model based on nonlocal information was proposed to removenoise and preserve the edges. Nonlocal gradient is introduced to detect the edges and adjust the speed of the diffusion inside and outside the smooth region. Oneside scheme is taken for calculating the nonlocal gradient norm to preserve edge information in discret situation. Theorical analysis and numerical results demonstratethe performance of the new method.

Key words: image denoising, nonlocal information, partial differ ential equation

中图分类号: 

  • O175.2