Abstract: This paper proposes a super-resolution reconstruction algorithm based on multiple low-resolution images, which is suitable for inner independence model or static state model. Various weights and regularization parameters are assigned to each low-resolution image, and weights and regularization parameters are updated in each iteration process. The optimal solutions are obtained with relaxation iteration method, so that the image is reconstructed. Experimental result shows that the effect of reconstruction with the algorithm is better than those of cubic B-spline interpolation and traditional MAP estimation method.

Key words: self-adaptive, regularization, super-resolution, relaxation iteration

摘要： 提出一种适用于内部独立运动、静态等多种模型的、基于多幅低分辨率图像的超分辨率重构算法。该算法赋予各低分辨率图像不同的权重和正则化参数，在每次迭代时对权重系数和正则化参数进行更新，采用松弛迭代法得到最优解，从而获得重构图像。实验结果表明，该算法获得的重构图像效果优于双三次B样条插值法和传统的最大后验概率估计方法。

关键词: 自适应, 正则化, 超分辨率, 松弛迭代

CLC Number: 

• TP301.6