基于l0高阶全变差的运动模糊图像盲复原算法

Blind recovery algorithm of motion blurred image based on l0 norm with high order total variation

为了改善运动模糊图像复原的效果,构造了基于l0高阶全变差的盲复原数学模型。该模型在l0稀疏先验的基础上,引入了可降低振铃效应的高阶全变差正则化约束项,用以解决低阶全变差造成的虚假边缘问题。盲复原算法首先用半二次分裂对数学模型求解,然后根据先验信息对模型解约束,多次迭代得到中间潜像、模糊核,最后采用非盲从复原算法恢复图像。实验结果表明:提出的方法可突出边缘结构特征,有效改善盲复原图像的效果。

In order to improve the restoration effect of motion fuzzy images,a blind restoration mathematical model based on high-order total variation is constructed.Based on sparse priori,the regularization constraint term of high-order total variation which can reduce the ringing effect is introduced to solve the false edge problem caused by low-order total variation.The blind restoration algorithm first solves the mathematical model by semi-quadratic splitting,then constrains the solution of the model according to prior information,and obtains the intermediate latent image and fuzzy kernel through multiple iterations.Finally,the non-blind restoration algorithm is used to restore the image.Experimental results show that the proposed method can highlight the edge structure features and improve the blind restoration image effect effectively.