中文摘要：

图像超分辨率重建是一种将多幅低分辨率图像合成为高分辨率图像的技术。当前的超分辨重建技术主要分为图像配准和超分辨率重建2个步骤，提出一种基于最大后验概率的图像超分辨率重建算法，将这2个步骤合二为一；与此同时，为了解决配准参数以及点扩展函数估计值的不精确性问题，在每一幅低分辨率图像代价函数的残差项引入了自适应加权系数并随之给出了迭代算法的总体框架。实验表明，该算法在收敛性和精确性上都达到了较好的效果。

英文摘要：

Super-resolution （SR） reconstruction produces a high-resolution （HR） image from multiple low-resolution （LR） images. Current SR techniques are commonly performed in two disjoint stages named image registration and image reconstruction. In this paper, we propose a new method based on maximum a posteriori （MAP） to joint the two stages for SR reconstruction. In order to solve the inaccurate problems of the registration parameters and the point spread function in the method, our algorithm introduces a self-adapting weight coefficient in the residual norm of the cost function of each LR frame. Then an iterative scheme is developed to get more accurate image for SR reconstruction. Our experimental results using both real and synthetic data show the effectiveness of the proposed algorithm.