紧框架分析模型下的模糊图像盲复原

发布时间：2019-05-29 07:19

【摘要】：在基于稀疏表示模型的图像盲复原问题中,模糊核估计与稀疏模型的选取是影响盲复原性能的两个关键因素。针对传统基于稀疏表示盲复原方法的不足,本文提出一种基于紧框架分析模型的图像盲复原方法。该方法将盲复原问题分裂为两个迭代的子问题,分别是基于梯度图像的模糊核估计与基于紧框架分析模型的非盲图像复原。在核估计问题中,提出同时约束核稀疏性及一阶微分平滑特性,进一步提高了核估计精度。在紧框架非盲图像复原问题中,提出一种基于Moreau envelope函数的数值计算方法,有效地解决紧框架复原模型的不可微和不可分离性。实验结果表明,本文复原方法在图像细节恢复与客观评价指标方面均优于传统复原算法。
[Abstract]:In the problem of blind image restoration based on sparse representation model, fuzzy kernel estimation and sparse model selection are the two key factors that affect the performance of blind restoration. In order to overcome the shortcomings of the traditional blind restoration method based on sparse representation, an image blind restoration method based on compact frame analysis model is proposed in this paper. In this method, the blind restoration problem is divided into two iterative subproblems, which are fuzzy kernel estimation based on gradient image and non-blind image restoration based on compact frame analysis model. In the kernel estimation problem, the kernel sparsity and the first order differential smoothing characteristics are proposed at the same time, which further improves the accuracy of the kernel estimation. In the problem of non-blind image restoration of tight frame, a numerical calculation method based on Moreau envelope function is proposed to effectively solve the indifferentiability and inseparability of compact frame restoration model. The experimental results show that the proposed restoration method is superior to the traditional restoration algorithm in image detail restoration and objective evaluation index.
【作者单位】：哈尔滨理工大学计算机科学与技术学院博士后科研流动站;
【基金】：国家自然科学基金项目(61501147) 中国博士后基金项目(2016M601438) 黑龙江省自然科学基金项目(F2015040) 黑龙江省博士后基金项目(LBH-Z15099)
【分类号】：TP391.41

【参考文献】