计算机工程与应用 ›› 2012, Vol. 48 ›› Issue (28): 197-200.

• 图形、图像、模式识别 • 上一篇    下一篇

基于图像局部最大中值差和梯度的非线性扩散

周  圆1,马文华2,徐海水3   

  1. 1.广东工业大学 计算机学院,广州 510006
    2.广东外语外贸大学 信息科学技术学院,广州 510006
    3.广东工业大学 网络信息与现代教育技术中心,广州 510006
  • 出版日期:2012-10-01 发布日期:2012-09-29

Nonlinear isotropic diffusion based on local maximum absolute median difference and gradient

ZHOU Yuan1, MA Wenhua2, XU Haishui3   

  1. 1.Faculty of Computer, Guangdong University of Technology, Guangzhou 510006, China
    2.School of Informatics, Guangdong University of Foreign Studies, Guangzhou 510006, China
    3.Center of Campus Network Information & Modem Educational Technology, Guangdong University of Technology, Guangzhou 510006, China
  • Online:2012-10-01 Published:2012-09-29

摘要: 分析了图像平滑中几种基于非线性扩散模型存在的问题,提出一种改进的非线性扩散方法。该方法用图像局部最大中值差和像素的梯度幅值联合度量图像的不平滑度,由此控制偏微分方程(PDE)的扩散行为,达到既去除噪声又保持图像边缘的目的。实验结果表明,用该方法既可以去除图像噪声又可以保持图像边缘,并且收敛速度较快。

关键词: 偏微分方程, 非线性扩散, 图像平滑

Abstract: This paper discusses problems that exist in some nonlinear diffusion model for image smoothing, and then proposes an improved nonlinear diffusion based on visual masking effect and magnitude of gradient. As a result of making use of local maximum absolute median difference and magnitude of image gradient, this method achieves better result in noise-removed and edge preservation, with faster convergence. Experimental result shows that this method is able to achieve a good result.

Key words: partial differential equation, nonlinear diffusion, image smoothing