计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (13): 170-172.

• 图形、图像、模式识别 • 上一篇    下一篇

一种改进的各向异性扩散图像去噪方法

陈一虎,叶正麟   

  1. 西北工业大学 理学院,西安 710072
  • 收稿日期:2007-08-21 修回日期:2007-11-19 出版日期:2008-05-01 发布日期:2008-05-01
  • 通讯作者: 陈一虎

Improved anisotropic diffusion image denoising method

CHEN Yi-hu,YE Zheng-lin   

  1. School of Science,Northwestern Polytechnical University,Xi’an 710072,China
  • Received:2007-08-21 Revised:2007-11-19 Online:2008-05-01 Published:2008-05-01
  • Contact: CHEN Yi-hu

摘要: 研究了基于图像特征方向的正交坐标系,分析了在此框架下的各向异性扩散图像去噪原理。然后根据人类视觉系统的一些特性提出了一种改进的各向异性扩散方法。该方法避免了各向异性扩散方程的不适定问题。实验结果表明,该方法在噪声消除和边缘保留方面能获得较好的效果。

关键词: 各向异性扩散, 图像特征方向, 人类视觉系统, 图像去噪

Abstract: This paper studies orthogonal coordination based on the directions of the image feature,then analyzes anisotropic diffusion image denoising method under this framework,and proposes an improved anisotropic diffusion method.The method avoids ill-posed problems of anisotropic diffusion equation.Experiments show that this method can achieve good effects in noise-removing and margin-preserving.

Key words: anisotropic diffusion, directions of the image feature, human visual system, image denoising