基于图像配准的POCS超分辨率图像重构

POCS Super-Resolution Image Reconstruction Based on Image Registration

图像重构是数字图像处理的一个重要的分支 ,根据图像序列进行重构的高分辨率图像在多种应用邻域得以应用。该文简要介绍了图像成像原理以及降阶模型 ,并提出在图像配准的基础上 ,对图像序列采用POCS方法高分辨率重构 ,将图像校准算法和POCS方法有机地结合在一起 ,同时给出了其中详细的算法和实现过程。实验仿真结果表明该算法当图像间平移量小于 10个象素、旋转角小于 5时收敛 ,且运算量小 ,收敛速度快 。

Image Reconstruction is an important part in the field of digital image processing, and high resolution image reconstructed from image sequences is applied to several fields. In the paper, imaging theory and degraded model are introduced, and a algorithm based on image registration is proposed with which algorithms and realization are given in detail, POCS technique is used to reconstruct high resolution image from image sequences, incorporating image registration technique and POCS technique. Experimental results demonstrate, when shift less than 10 pixels and rotation angle less than 5 between images, the algorithm converges, and it is characterized by low computation complexity, fast convergence and good reconstruction effect.