基于改进混合范数的图像重构算法

摘要

压缩感知理论利用图像稀疏表示的先验知识,通过少量的测量值精确地恢复出原始图像信号.DTCWT凭借其更好表达图像边缘特征的特点,结合贪婪算法和凸优化算法的优点,在双树复数小波变换为稀疏基,局部哈达玛矩阵为观测矩阵的基础上提出改进的快速二步迭代混合范数算法.该算法的图像重构质量优于FIST算法以及IST算法,实验表明改进的混合范数的图像重构算法具有更好的图像重构质量和重构速度.%The compressed sensing theory utilizes a priori knowledge of sparse images to accurately recover the original image signal with a small amount of measured values.DTCWT,by virtue of its ability to express the characteristics of image edge feature,combined with the advantages of greedy algorithm and convex optimization algorithm,the paper proposes an improved fast two-step method based on double tree complex wavelet transform as sparse base and local Hadamard matrix as observation matrix iterative mixed norm algorithm.The image reconstruction quality of this algorithm is better than FIST algorithm and IST algorithm.The experimental results show that the image reconstruction algorithm with improved mixed norm has better image reconstruction quality and reconstruction speed.