基于稀疏表示的自适应图像超分辨率重建算法

摘要

Conduct research on the existence and uniqueness of sparse resolution in super-resolution image reconstruction and the relationship of edge feature and smoothing noise of super-resolution image. In order to solve these problems, this paper proposed the mean of adaptive regularization. Combining with joint over-complete dictionary, it realized dynamically adjust the regularization parameter. Through super-resolution image reconstruction experiment, it demonstrates that this algorithm is working efficiently and balances edge features and smoothing noise well, which has high PSNR compared with the traditional algorithm.%针对图像高分辨率重建过程中稀疏解的存在性和唯一性问题以及超分辨率图像的边缘特征和平滑噪声的关系进行了研究,提出了局部正则化参数自适应选取的方法.结合联合构造字典的算法,在重建过程中动态调整正则化参数.通过对图像的超分辨率实验证明,改进的算法具有较高的可行性,能有效平衡超分辨率图像的边缘特征和平滑噪声两者的关系,与传统的超分辨率重建算法相比,有更高的峰值信噪比.