Abstract: Proposes an image super-resolution algorithm that takes into account inaccurate estimates of the registration parameters and the point spread function.There inaccurate estimates result in different noise level for each Low-Resolution（LR） frames.In the proposed algorithm，the LR frames are adaptively weighted according to their reliability.Using gradient descent optimization procedure，the weights are iteratively updated along with the high-resolution image.Experiment results indicate that this method is better than the conventional MAP methods on objective statistical analysis.

Key words: super-resolution, image restoration, registration parameters

摘要： 提出一种图像超分辨率复原算法，该算法考虑了运动矢量和点扩散函数估计不精确的情况。由于估计的不精确造成每帧低分辨率图像具有不同的噪声，算法根据各帧低分辨率图像的可靠性，相应地给它们赋予不同的权值，应用梯度下降法不断对权值及高分辨率图像进行更新。实验结果表明该算法较传统的MAP算法具有更好的复原效果。

关键词: 超分辨率, 图像复原, 运动参数