计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (36): 200-201.

• 工程与应用 • 上一篇    下一篇

基于整体最小二乘法的高分辨率图像重建算法

解 凯1,吕妍昱2   

  1. 1.北京印刷学院 信息与机电工程学院,北京 102600
    2.北京印刷学院 教学技术与计算中心,北京 102600
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-12-21 发布日期:2007-12-21
  • 通讯作者: 解 凯

Total least squares based high-resolution image reconstruction algorithm

XIE Kai1,LV Yan-yu2   

  1. 1.School of Information and Mechanical Electronic Engineering,Beijing Institute of Graphic Commnunication,Beijing 102600,China
    2.Teaching Technology & Internet Center,Beijing Institute of Graphic Communication,Beijing 102600,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-12-21 Published:2007-12-21
  • Contact: XIE Kai

摘要: 提出一种用于高分辨率图像重建的整体最小二乘算法。在现有多数重建算法中,假设系统矩阵是精确的而误差主要源于采样图像,但实际上抖动误差也出现在系统矩阵中。该方法能同时最小化这两种误差,采用基于正则化的Rayleigh商来光滑解,用共轭梯度算法来迭代求解该正则化Rayleigh商的最小化函数。实验证明该方法对于抖动系统矩阵是稳定和精确的。

关键词: Rayleigh商, 整体最小二乘法, 高分辨率图像

Abstract: The paper presents a High-Resolution(HR) image reconstruction algorithm based on total least squares approach.Many HR algorithms assume that system matrix is accurate and errors are from the sampled Low-Resolution(LR) images.In practice,errors also occur in the system matrix.The paper proposes the TLS approach that minimizes the two kinds of errors.A Rayleigh Quotient(RQ) formulation based on regularization is used to enforce a smooth solution.Then a conjugate gradient algorithm is used to minimize the modified RQ function.Experiment results show that the method is stable and accurate for the perturbation system matrix.

Key words: Rayleigh quotient, total least squares, high-resolution image