Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (32): 180-182.DOI: 10.3778/j.issn.1002-8331.2008.32.053

• 图形、图像、模式识别 • Previous Articles     Next Articles

Iterative projected regularizing methods for digital image restoration

MA Qing-hua,JIN Wei-ping   

  1. College of Art and Science of Beijing Union University,Beijing 100083,China
  • Received:2007-12-10 Revised:2008-03-18 Online:2008-11-11 Published:2008-11-11
  • Contact: MA Qing-hua

数字图像恢复的迭代投影正则化方法研究

马青华,靳伟萍   

  1. 北京联合大学 应用文理学院,北京 100083
  • 通讯作者: 马青华

Abstract: The purpose of image restoration aims at reconstructing the original input image from the corrupted degraded image,such that it approaches the perfect image that is deblurred and has no noise.The iterative regularization methods is studied,particularly the gradient methods including Least Squares Error(LSE) method,steepest descent method and Conjugate Gradient(CG)method.Particularly a projected CG method is proposed because of the physical meaning of the practical problem.Numerical simulations are performed on atmospheric blurred remote sensing image.

Key words: image restoration, Least Squares Error(LSE), regularizing methods, projection

摘要: 图像恢复的目的是对模糊的图像进行处理,使它趋向于复原的或没有噪声影响的理想图像。研究图像恢复的迭代正则化方法,主要研究梯度法,特别是提出了求图像恢复的投影共轭梯度法。针对受大气扰动影响的遥感图像,根据梯度法使用MATLAB进行编程,对模糊且带噪音的图像进行恢复,并与经典的最小二乘法进行比较。

关键词: 图像恢复, 最小均方误差, 正则化, 投影