【摘要】 超分辨率图像重构是将多帧低分辨率图像重构成一幅高分辨率图像的过程。由于其求解是一大型病态求逆问题,计算量随着放大倍数的增加而急剧上升,如何降低计算复杂度是超分辨率成像所面临的一个急需解决的课题。提出了一个基于POCS的高分辨率图像重构的快速算法。其原理是利用各低分辨率图像之间位移的关系将所有的低分辨率图像进行重组,然后对每个组进行POCS超分辨图象重构。实验结果表明,该快速算法较大地提高了超分辨图像重构的速度。更多还原

【Abstract】 Super-resolution image reconstruction produces a high-resolution image from a set of low-resolution images. Since super-resolution image reconstruction is a large scale ill-conditioned problem and the computational complexity increases dramatically as magnification increases. In this paper, a new fast algorithm is proposed, to improve the computational efficiency of POCS-based super-resolution algorithms. The principle of the fast algorithm is using the relation of translation of every low-resolution images to group them into several groups and apply conventional POCS algorithm to each group. The experimental tests show that with respect to conventional POCS algorithm, the fast algorithm has a significant increase of efficiency.更多还原