J4 ›› 2012, Vol. 39 ›› Issue (6): 167-169+186.doi: 10.3969/j.issn.1001-2400.2012.06.027

• 研究论文 • 上一篇    下一篇

一种基于稀疏正则化的图像盲复原方法

王书振;邹子健;李莉;张小平   

  1. (西安电子科技大学 计算机学院,陕西 西安  710071)
  • 收稿日期:2012-05-17 出版日期:2012-12-20 发布日期:2013-01-17
  • 通讯作者: 王书振
  • 作者简介:王书振(1978-),男,副教授,博士,E-mail: shuzhenwang@xidian.edu.cn.
  • 基金资助:

    国家自然科学青年基金资助项目(61100156);中央高校基本科研业务费专项资金资助项目(K50511030007);国家自然科学基金资助项目(61070143)

Method for blind image restoration based on sparse regularization

WANG Shuzhen;ZOU Zijian;LI Li;ZHANG Xiaoping   

  1. (School of Computer Science and Technology, Xidian Univ., Xi'an  710071, China)
  • Received:2012-05-17 Online:2012-12-20 Published:2013-01-17
  • Contact: WANG Shuzhen

摘要:

在图像盲反卷积的过程中,最主要的难点是缺少点扩散函数的足够信息而导致的病态问题.解决此问题可以通过对原始图像和点扩散函数同时进行正则化约束.为了在图像复原过程中得到惟一、稳定的解,并保证图像恢复结果的有效性,提出了一种具有尺度不变性和稀疏性的正则化函数,并通过两组对比实验例证了利用该函数的图像盲复原算法具有良好的鲁棒性和收敛稳定性.

关键词: 图像盲复原, 稀疏性表示, 解卷积

Abstract:

In the process of image blind deconvolution, the main obstacle is the lack of enough information about the point spread function (PSF), which leads to the ill-posed problem. To solve this problem, we can give regularization constraints on the original image and PSF simultaneously. In order to gain the stable and unique solution and guarantee the effectiveness of the resulting image restoration, this paper uses a scale invariant and sparse regularization function, and experiments are conducted to verify that our image blind recovery algorithm is robust and has stable convergence.

Key words: image blind restoration, sparse representation, deconvolution