Super-resolution Image Reconstruction Methods And Applications Based On Compressive Sensing

In this paper, a method of image super-resolution reconstruction based on compressed sensing is studied. In order to solve the block artifacts of the reconstructed image, the post-processing procedure of an adaptive TV denoising and an iterative back projection(IBP), are introduced after the super-resolution reconstruction method based on block compressed sensing. After the whole procedure, the quality namely PNSR of the reconstructed image is higher compared with the classical interpolation algorithm.In the post-treatment procedure, the block artifacts of the reconstructed image are treated as noise and dealt with an adaptive TV denoising method. Although the image quality improves, but the reconstructed image is fuzzy after denoising, so the iterative back-projection, IBP, is used to optimize the image quality.In addition, in the block compressed sensing of super-resolution reconstruction process, a new algorithm is developed through combining Orthogonal Matching Pursuit(OMP) algorithm with the steepest descent method.Simulation experiment results show that the new algorithm achieved better image quality and the reconstruction speed in image of the same dimension reconstruction process. But the improved algorithm has same effect with OMP algorithm when applied in reconstruction of the super-resolution of image.Finally, the proposed post-processing procedure is used to super-resolution reconstruction of color images. Compared with the classical interpolation algorithm, it also has better reconstruction quality.