提出一种参数自适应的在线字典学习图像超分辨率重建算法。在经典的稀疏表示算法框架下,运用在线字典学习方法来提高字典学习的精度。通过参数自适应方法灵活调整稀疏重建阶段的正则化参数,并依据每个图像块的特点自适应确定正则化参数,以此克服人为设定参数的单一性和非最佳参数值的缺点。实验结果表明,与传统算法相比,所提算法可有效降低测试图像对训练图像集的依赖程度,同时克服图像在重建过程中存在的局部模糊或失真,进一步提高重建图像的质量。

In this paper, an super-resolution imaging reconstruction algorithm based on the parametric adaptive online dictionary learning (ODL) is proposed. Under the framework of the classical sparse representation algorithm, the ODL method is used to improve the accuracy of dictionary learning. Furthermore, the regularization parameters in the sparse reconstruction stage are flexibly adjusted using the parameter adaptive method, so that the regularization parameters can be adaptively determined based on the characteristics of each image block, overcoming the disadvantages of the singularity and incompatibility of the artificially set parameters. Results show that compared with the traditional algorithm, the proposed algorithm can reduce the dependence of test images on the training image set, overcome the local blur or distortion in the reconstruction process, and improve the quality of the reconstructed image.