Abstract: Image superresolution(SR) reconstruction technique is how to produce a clearly high resolution(HR) image from the information of one or several low resolution(LR) images, and it has been received increasing attention from the image processing community. In the paper, the fundamental principles and mathematical models of SR technology based on reconstruction were described firstly, and the history and stateofart of image SR reconstruction were stated briefly according to the classification between the frequency domain method and the space domain method. Secondly, advantages and defects of different methods were analyzed and summarized systematically. Finally, the further research directions for image SR technique of reconstructing based the approach were proposed.

Key words: Superresolution, Image reconstruction, Comparative analysis, Image processing, Regularization　

摘要： 图像超分辨率(superresolution,SR)重建技术是利用一帧或多帧低分辨率(lowresolution,LR)图像
的信息来重建一帧清晰的高分辨率(highresolution,HR)图像的技术,是图像处理中的研究热点。介绍了基于重建
方法的图像SR技术的基本原理及数学模型,以频域方法和空域方法作为分类依据,分别阐述了图像SR重建技术的
经典方法和最新进展,并对各类算法的优缺点进行了系统的分析和总结,最后指出了基于重建方法的图像SR技术
的研究方向。

关键词: 超分辨率 , 图像重建 , 对比分析 ,  图像处理 , 正则化