摘要图像超分辨率重建技术是从单幅或者多幅低分辨率图像重建得到高分辨率图像的技术，其基本的思想是采纳信号处理的方法，恢复成像的重建过程中丢失的高频率信息。目前，图像超分辨率重建算法主要分为三大类：基于插值的图像超分辨率算法，基于多帧重建的图像超分辨率算法以及基于学习的图像超分辨率算法。本文首先对基于学习的图像超分辨率重建算法和稀疏表示理论进行了介绍，其次对本文中将要使用的重建算法以及弹性网约束进行了讨论。最后使用该方法进行了多组仿真实验，并给出了结果以及总结。关键词：超分辨率重建，基于学习的方法，稀疏表示，弹性网约束Abstract Image super-resolution technique is to get high-resolution images from a single low-resolution image or multiple low-resolution image, the basic idea is to use the signal processing method, to recovery the high frequency information which is lost in the imaging reconstruction process. Currently, the super-resolution image reconstruction algorithm is mainly divided into three categories: Image super-resolution algorithms based on interpolation, image super-resolution algorithms based on multi-frame and image super-resolution reconstruction algorithms based on learning. In this thesis, firstly, image super-resolution reconstruction algorithms based on learning and sparse representation theories were introduced, then the reconstruction algorithm used herein as well as elastic net were studied. Finally, multiple sets of simulation experiments were performed, and the results and summary were given.Keywords: Super-resolution reconstruction, Learning-based approach, Sparse representation, Elastic net目录1 前言11.1 课题背景 11.2 超分辨率重建技术及其实际应用领域 21.3 本文的主要工作及内容安排82 稀疏表示与字典学习理论102.1 稀疏表示理论102.2 稀疏表示的字典构建方法122.3 弹性网约束概述 162.4 本章小结 183 基于稀疏表示的单幅图像超分辨率重建 193.1 算法流程概述193.2 重建流程概述203.3 图像质量评价213.4 本章小结 244 实...
