图像退化仿真系统的设计与实现

超分辨率图像重建效果是由参与重建的各个低分辨率图像的质量、数量、图像的相对关系等因素及重建算法决定的。因此在超分辨率重建研究中,往往需要用一些已知的图像来研究算法的效果。现实中,要按算法实验的要求获取一定数量的同一场景的低分辨率图像是很困难的。本文设计并实现了一种基于MATLAB语言的图像退化仿真系统,该系统可以用一张高分辨率图像作为图像源,按实验要求,经过图像形变(包括旋转、亚像素平移、模糊化、降采样和加噪)采样得到一组低分辨率图像;利用这组低分辨率图像就可以研究图像重建的效果及方法的可行性等。实验表明,该图像退化仿真系统是可行的,由于有原图像,还可以对重建效果进行很好的评价。

Designing and implementation of the simulation system of image degradation

Effect of super-resolution image reconstruction is decided by the quality of each low-resolution image, quantity of images used, the relative relationship between the images, and the reconstruction algorithm. So in the super-resolution reconstruction study, some known images are often used to study the effect of algorithm. In reality, it is difficult to obtain a number of low-resolution images of the same scene according to the experimental requirements. This paper designs and implements a image degradation method based on MATLAB simulation system. It uses a high resolution an image as image source, through the image deformation （including rotation and sub-pixel translation, fuzzy, down-sampling, plus noise） to sample a group of low-resolution images. This group of low-resolution images can then be used to study about the effect of image reconstruction. Simulation results show that the image system is feasible, with the original image, the effect of reconstruction can also be well evaluated.