基于小波稳健的超分辨率图像重建算法  

A robust regularization super-resolution image reconstructing algorithm based on wavelet

作　　者：李淑静[1] 邵峰晶[1] 

机构地区：[1]青岛大学信息工程学院,山东青岛266071

出　　处：《微型机与应用》2007年第S1期91-93,共3页Microcomputer & Its Applications

基　　金：国家高技术研究发展计划(863计划)基金(2003AA1Z1110);国家科技攻关引导项目基金(2001BA105C)资助。

摘　　要：针对超分辨率图像重建算法多存在计算量大、收敛稳定性不高且收敛慢的问题,提出一种基于小波稳健的正则化超分辨率图像重建算法。该算法利用小波变换生成初始图像,对重建图像的质量有明显提高。采用1-范式(L1)度量正则项,增强了算法的稳健性,通过导入自适应的正则参数提高了算法的效率。经仿真实验证明了算法的有效性。In order to solve the problem of the rebuild algorithm of super resolution images, which have high computing complexity, slow convergent velocity and unstable convergence, this article presents a robust and regularizing rebuilding algorithm of super resolution. This algorithm creates initial images via wavelet, and improves the quality of rebuilt images. Measuring regularization using L1 norm, makes the algorithm robust. It improves the efficiency of the algorithm with importing adaptive regularization para...

关 键 词：超分辨率 小波变换 正则化 凸集投影 图像融合 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]