一种基于自适应正则化的图像超分辨率重建模型

Image Super-resolution Reconstruction Algorithm Based on Auto-adaptive Regularization

摘要:

为了解决图像超分辨率重建中稀疏系数解的不精确问题,提出了一种自适应正则化级联稀疏矩阵的超分辨率重建算法。根据图像自身的特性,采用自适应正则化项对图像局部进行处理,实现图像的局部约束,构建基于自适应正则化的稀疏矩阵函数。另外,为了提高图像的可清晰性,采用基于全局约束的退化模型改进处理结构。测试结果表明,与其他常用算法相比,提出的自适应正则化的图像超分辨率重建算法能够构建更清晰的超分辨率图像。

Abstract:

In order to solve the imprecise problem of the sparse coefficient solution in image super-resolution reconstruction, a super-resolution reconstruction algorithm based on adaptive regularized cascade sparse matrix was proposed. According to the characteristics of the image itself, adaptive regularization items were used to process the local image, so as to realize the local constraints of the image, and build a sparse matrix function based on adaptive regularization. In addition, in order to improve the clarity of the image, a degradation model based on global constraints was adopted to improve the process structure. Test results show that the proposed algorithm can construct clearer super-resolution images compared with other commonly used algorithms.