基于最小二乘准则的模糊估计和图像复原

被引:2
作者
卿粼波
何小海
陶青川
吕成淮
张菊
孙贵凡
机构
[1] 四川大学电子信息学院图像信息研究所
关键词
计算光学显微成像; 最小二乘; 模糊估计; 图像复原;
D O I
10.15961/j.jsuese.2008.02.027
中图分类号
TP391.41 [];
学科分类号
080203 ;
摘要
在计算光学显微成像技术中,点扩展函数往往是未知的,且不易获取,从而给图像复原带来很大困难。基于最小二乘准则和最优化理论,提出了利用变尺度法的三维点扩展函数参数估计算法;针对传统EM算法存在复原效果细节丢失严重等问题,提出最小二乘共轭梯度三维图像复原算法。算法在点扩展函数参数估计和求解真实图像之间进行交替迭代,从而得到图像的最优估计。实验表明,新算法在较短时间内,能够较准确地估计出点扩展函数参数,并得到较好的复原结果。
引用
收藏
页码:129 / 133
页数:5
相关论文
共 12 条
[1]  
高斯模糊图像的盲复原[D]. 顾亚芳.东南大学 2006
[2]  
A new method forsolving line equations with large sparse symmetric and Indef-inite coefficients matrix. Wang Jinming,Xie Dexin,Bai Baodong. IEEE Transactions on Magnetics . 2004
[3]  
A statistical foundation for the useof the conjugate gradient method in deconvolution. Elias E D,Pattipati K R. Pro-ceedings of the 35thConference on Decision and Control . 1996
[4]  
Steepest descent,CG and iterativeregularization of Ill-posed problems. Nagy J G,Palmer K M. Tidskrift for Infor-mation Behandling . 2003
[5]  
Regularization methods for large-scale problems. Hanke M,Hansen P C. Sruveys on Mathematics for Industry . 1993
[6]  
Iterative regularization techniques in imagereconstruction. Martin Hanke. Proceedings of the Conference Mathe-matical Methods in Inverse Problems for Partial DifferentialEquations . 1998
[7]  
最优化计算原理与算法程序设计[M]. 国防科技大学出版社 , 粟塔山等编著, 2001
[8]  
Steepest Descent, CG, and Iterative Regularization of Ill-Posed Problems[J] . J. G. Nagy,K. M. Palmer. &nbspBit Numerical Mathematics . 2003 (5)
[9]  
Fast regularization techniquefor expectation maximization algorithm for computational op-tical sectioning microscopy. Conchello J A,,McNally J G. Procof the InternationalSociety for Optical Engineering . 1996
[10]  
Numeri-cal recipes in C:The art of scientific computing. Press W H,Teukolsky S A,Vetterling W T,et al. . 1992