分数域图像混合噪声盲复原方法  

A New Method of Mixed Noise Image Restoration in Fractional Fourier Transform Domain

作　　者：刘莉[1,2] 谭吉春[3] 

机构地区：[1]国防科技大学光电学院,湖南长沙410073 [2]长沙学院电子与通信工程系,湖南长沙410003 [3]国防科技大学理学院,湖南长沙410073

出　　处：《国防科技大学学报》2011年第3期52-55,98,共5页Journal of National University of Defense Technology

摘　　要：提出了一种在分数傅里叶变换域中对图像进行复原的新方法。对于模糊且含有高斯噪声和啁啾噪声的图像进行分数傅里叶变换,利用各种噪声在分数域中的性质估计出复原图像的分数谱,再通过逆分数傅里叶变换返回空域,获得复原图像。在图像受到多种因素导致退化的情况下,以及在无任何先验知识的前提下,此方法可以简化复原过程,进行图像盲复原。A new method of image restoration based in Fractional Fourier Transform （FRT） domain is presented. The blurred image including higher intensity Gaussian noise and chirp noise is transformed into FRT domain. Only when the noise FRT spectrum is well estimated, the required image FRT spectrum can be well separated and return to the space domain with the inverse ransformation. As a result, we find that the method simplifies the process of restoration and the blind restoration will be obtained when we know nothing about the detwadafion process affected by various factors.

关 键 词：分数傅里叶变换 噪声谱 图像复原 

分 类 号：O438.1[机械工程—光学工程]