一种改进的NAS-RIF盲图像复原算法  

An Improved NAS-RIF Blind Image Restoration Algorithm

作　　者：杨彦[1] 罗代升[1] 陶青川[1] 

机构地区：[1]四川大学电子信息学院,四川成都610064

出　　处：《信息与电子工程》2003年第4期260-264,共5页information and electronic engineering

基　　金：国家自然科学基金资助项目(30070228)

摘　　要：在DeepaKundur等人提出的NAS-RIF(Nonnegativityandsupportconstraintsrecursiveinversefiltering)盲图像复原算法的基础上,提出改进方法。首先对非均匀背景图像进行背景校正,以扩大原算法的适应范围;接着改进求取支撑域的算法,以提高图像的复原效果;最后对求取梯度值的算法作了改进,以提高原算法的执行效率。由实验结果可看出,改进的NAS-RIF算法比原算法具有更好的图像复原性。Based on NAS-RIF algorithm for blind image restoration proposed by Deepa Kundur, an improved method is proposed in this paper.In this method, background rectification is developed for non-uniform background images to expand the applied scope of the original algorithm.Two algorithms for calculation of support region and image gradient are also improved to obtain higher execution efficiency and image restoration effects.Computer simulations show that this improved NAS-RIF algorithm has better image restoration.

关 键 词：信息处理技 背景校正 支撑域 梯度运算 图像分割 NAS-RIF盲图像复原算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]