计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (17): 172-174.

• 图形、图像、模式识别 • 上一篇    下一篇

一种新的线性混合滤波器

罗晓军1,3,李 兵1,LIU Quansheng1,2,徐家伟1   

  1. 1.长沙理工大学 数学与计算科学学院,长沙 410076
    2.Universite de Bretagne-Sud,Campus de Tohannic,56017 Vannes,France
    3.商丘职业技术学院 计算机系,河南 商丘 476000
  • 收稿日期:2007-12-28 修回日期:2008-03-21 出版日期:2008-06-11 发布日期:2008-06-11
  • 通讯作者: 罗晓军

New linear mixed filter

LUO Xiao-jun1,3,LI Bing1,LIU Quan-sheng1,2,XU Jia-wei1   

  1. 1.School of Math. and Computing Science,Changsha Univ. of Science and Technology,Changsha 410076,China
    2.Universite de Bretagne-Sud,Campus de Tohannic,56017,Vannes,France
    3.Department of Computer,Shangqiu Vocational and Technical College,Shangqiu,Henan 476000,China
  • Received:2007-12-28 Revised:2008-03-21 Online:2008-06-11 Published:2008-06-11
  • Contact: LUO Xiao-jun

摘要: 提出一种新的滤波器,称为线性混合滤波器(简称为LMF),它适用于恢复被一类混合噪音(即一致脉冲噪音与高斯噪音)污染的数字图像。当混合噪音强度在一定范围内变化时,它具有自动调节机制。与若干已知的同类滤波器相比,LMF的速度更快,具有简单而统一的控制参数计算公式,而不是关于参数的选取范围。而且从实验结果可见,用LMF时峰值信噪比得到提高,均方误差得到降低。

Abstract: A novel filter called Linear Mixed Filter(LMF) is proposed in this paper,which is suitable for digital images corrupted with mixed noises(impulse noise and Gaussian noise).It has an automatic adjustment mechanism when the noise change.Comparing with some similar filters,it needs fewer time to restore an image,with simple and unified control parameter formulas(not a range of them) and the peak signal to noise ratio for a number of denoised images are increased while the mean square errors are reduced by experiments.