超分辨率图像复原中的快速L-曲线估计  

Efficient Parameter Estimation with L-Curve in Super-Resolution Image Reconstruction

作　　者：盛琳阳[1] 解凯[2] 张田文[3] 

机构地区：[1]哈尔滨师范大学计算机系,黑龙江哈尔滨150080 [2]北京印刷学院信息与机电学院,北京102600 [3]哈尔滨工业大学计算机科学技术学院,黑龙江哈尔滨150001

出　　处：《小型微型计算机系统》2006年第6期1120-1123,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60075010)资助;黑龙江省教育厅科学基金项目(10551115)资助;哈尔滨师范大学科研基金项目(KM2005-11)资助

摘　　要：讨论了从一组低采样降质的视频图像重建超分辨率图像中未知参数的估计问题.使用L-Curve标准来估计正则化参数.然而,L-Curve的计算代价十分昂贵.它需要计算正则化近似解和残差的范式.为此提出一种基于Lanczos算法和Gauss积分理论的算法,在超分辨率图像重建中的参数估计中可以减少L-Curve的计算代价.this paper considers the estimation of the unknown parameter for the problem of reconstructing a super-resolution image from multiple under-sampled, shifted, and degraded frames. L-curve criterion is used to estimate the regularization param eters. However, the computation of the L-curve is quite costly for large problems. The determination of a point in the L-curve requires that both the norm of the regularized approximate solution and the norm of the correspondlngresidual vector are available. The paper proposes an approximate technique based on the Lanczos algorithm and Gauss quadrature theory, which can reduce computational complexity of the L-curve for the efficient parameter estimation in high-resolution image reconstruction.

关 键 词：L-CURVE GAUSS积分 超分辨率图像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]