一种改进的全变差盲图像复原方法

当点传播函数未知或不确知的情况下,从观察到的退化图像中复原原始图像的过程称为图像盲复原.传统的图像盲复原算法常采用最小均方误差作为复原效果的评判准则,但它很少考虑人类视觉心理,而图像最终都必须由人类的视觉系统来观测和解释.因此,本文提出一种新的基于人类视觉特性的图像盲复原算法:它采用交替最小化的结构,在模糊辨识阶段,采用全变差正则化算法;在复原阶段,采用基于Weber定律和全变差正则化相结合的算法.仿真实验表明,这种算法可在未知点扩展函数的情况下取得较好的复原效果.

A Modified Method for Total Variation Blind Image Restoration

When Point Spread Function (PSF) is not known or only partially determined,restoration of the degraded images is called blind image restoration.MSE is used as criterion in the traditional blind image restoration methods,which considers little the influence of human vision psychology.But all images are eventually perceived and interpreted by the human visual system.So this paper proposes a new blind image restoration algorithm based on models of visual characteristic.An iterative scheme using alternating minimization is devised to recover the image and simultaneously identify the PSF.In the process of identification,total variation (TV) regularization method is used.Then in the process of restoration,Weberized TV restoration approaches is applied.The result of experiment shows that this restoration model produces better effect without known PSF.