空间移不变系统图像超分辨复原的快速解耦算法

超分辨率图像复原是当今一个重要的热门研究课题. 本文提出了一种基于全变差模型的超分辨率复原快速解耦算法. 利用半二次正则化思想, 提出了一个新的解耦TV (Total variation)模型. 利用交替最小化方法和线性空间不变模糊的性质将上采样融合、去模糊和去噪分步进行. 算法中对上采样融合采用非迭代的直接计算方法; 去模糊过程采用基于变换的预处理共轭梯度迭代算法, 而去噪过程采用了子空间投影方法. 本文算法降低了算法复杂度; 超分辨率重建图像在去除噪声的同时, 不仅能够保证图像平坦区域的保真度, 较好地抑制阶梯效应的产生, 而且能够保持图像中边缘等重要几何结构的清晰度.

A Fast Decoupling Algorithm for Image Super-resolution Reconstruction of Space-invariant System

Super-resolution reconstruction has been a very hot research topic currently. In this paper, we study a fast decoupling algorithm for TV-based (Total variation) image super-resolution reconstruction. A new TV-based decoupling model is proposed by utilizing the half-quadratic regularization approach. The treatment is separated into measurements fusion, deblurring and denoising by exploiting the alternating minimization algorithm and the property of the space-invariance blur. The fusion part is shown to be a very simple non-iterative algorithm. The deblurring part can be solved by transform-based preconditioning conjugate gradient method for improved convergence. The subspace projection method is employed to solve the denoising problem. The proposed algorithm reduces the computational complexity. Super-resolution experiment results demonstrate that the proposed algorithm is not only able to restrain the noise and preserve the important geometric structures in the image such as edges, but also able to maintain the fidelity of flat regions without producing the staircase effect.