运动模糊图像盲解卷积快速算法

廖永忠1; 蔡自兴1; 何湘华2

doi:10.3788/OPE.20132110.2688

您当前的位置：

首页 >

文章列表页 >

运动模糊图像盲解卷积快速算法

信息科学 | 更新时间：2020-08-12

- 运动模糊图像盲解卷积快速算法
- Fast algorithm for motion blurred image blind deconvolution
- 光学精密工程 2013年21卷第10期页码：2688-2695
- 作者机构：
  
  1. 湖南第一师范学院信息科学与工程系,湖南长沙,410205
  2. 中南大学信息科学与工程学院，湖南长沙 410081
- 作者简介：
- 基金信息：
  
  高速公路车辆能驾驶中关键科学问题研究()
- DOI：10.3788/OPE.20132110.2688
  中图分类号： TP391.4
- 收稿日期：2013-03-04，
  
  修回日期：2013-05-14，
  
  网络出版日期：2012-10-19，
  
  纸质出版日期：2013-10-15
- 稿件说明：
移动端阅览
廖永忠蔡自兴何湘华. 运动模糊图像盲解卷积快速算法[J]. 光学精密工程, 2013,21(10): 2688-2695

LIAO Yong-zhong CAI Zi-xing HE Xiang-hua. Fast algorithm for motion blurred image blind deconvolution[J]. Editorial Office of Optics and Precision Engineering, 2013,21(10): 2688-2695
廖永忠蔡自兴何湘华. 运动模糊图像盲解卷积快速算法[J]. 光学精密工程, 2013,21(10): 2688-2695 DOI： 10.3788/OPE.20132110.2688.

LIAO Yong-zhong CAI Zi-xing HE Xiang-hua. Fast algorithm for motion blurred image blind deconvolution[J]. Editorial Office of Optics and Precision Engineering, 2013,21(10): 2688-2695 DOI： 10.3788/OPE.20132110.2688.

摘要

针对快速运动形成的图像模糊，提出了一种运动模糊图像盲解卷积算法。首先，对被噪声污染的频谱图像进行脊波增强；然后，采用一种新的基于Radon变换的鲁棒算法来确定模糊核函数，该算法在小模糊长度和低信噪比的条件下仍能准确地估计模糊核参数；确定模糊核函数后，采用基于hyper-laplacian先验的快速非盲解卷积算法来恢复模糊图像。实验结果证明，与基于机器学习的R.Fergus的算法相比较，本文算法在获得相近效果的前提下，计算时间从近30 min下降到40 s左右。该算法对合成运动模糊图像和实际相机运动的自然模糊图像都具有较好的恢复效果。

Abstract

For a blurred image caused by fast movement

a fast blind deconvolution algorithm for spatially-invariant motion blurred images was proposed. Firstly

the ridge wave of a frequency spectral image with noises was enhanced. Then a robust algorithm based on Ridgelet transform and Radon transform was used to estimate blur kernels in the frequency domain

by which the lengths and directions of motion blur kernels could be accurately estimated

even for small length parameters and blur images in low SNRs .Furthermore

a fast non-blind deconvolution method based on hyper-laplacian prior was used to restore blur images. Experimental results show that the proposed method can restore a 1 megapixel image in less 40 s. As compared with R. Fergus algorithm based on machine learning

the proposed algorithm reduces the computing time from 30 min to 40 s while keeps the comparable quality. Moreover

the algorithm is effective not only for the artificially blurred images

but also for the naturally blurred images (by camera movement) as well.

关键词

Keywords

references

KRAHMER F, LIN Y, MCADOO B, et al.. Blind image deconvolution: Motion blur estimation[J]. IMA Preprints Series, 2006: 2133-5.[2]DAI S, WU Y. Motion from blur [C].Computer Vision and Pattern Recognition(CVPR), 2008: 1-8.[3]MISKIN J. Ensemble Learning for Independent Component Analysis [D]. Cambridge:University of Cambridge.2000.[4]FERGUS R, SINGH B, HERTZMANN A,et al.. Removing camera shake from a single photograph[J].ACM Transactions on Graphics, 2006, 25(3): 787-794.[5]LEVIN A, WEISS Y, DURAND F, et al.. Efficient marginal likelihood optimization in blind deconvolution [C].Computer Vision and Pattern Recognition (CVPR), 2011: 2657-2664.[6]LEVIN A, WEISS Y, DURAND F, et al.. Understanding and evaluating blind deconvolution algorithms [C].Computer Vision and Pattern Recognition, 2009:1964-1971.[7]CHO T S, PARIS S, HORN B K P, et al.. Blur kernel estimation using the Radon transform [C].Computer Vision and Pattern Recognition (CVPR), 2011: 241-248.[8]DOBE M, MACHALA L, FRST T. Blurred image restoration: A fast method of finding the motion length and angle [J]. Digital Signal Processing, 2010, 20(6): 1677-1686.[9]SAKANO M, SUETAKE N, UCHINO E. A PSF estimation based on Hough transform concerning gradient vector for noisy and motion blurred images[J]. IEICE Transactions on Information and Systems, 2007, 90(1): 182-190.[10]DO M N, VETTERLI M. The finite ridgelet transform for image representation[J]. IEEE Transactions on Image Processing, 2003, 12(1): 16-28.[11]CANDES E J. Ridgelets: theory and applications[D]. Stanford:Stanford University, 1998.[12]STARCK J L, MURTAGH F, CANDES E J, et al.. Gray and color image contrast enhancement by the curvelet transform [J]. IEEE Transactions on Image Processing, 2003, 12(6): 706-717.[13]XU J, ZHANG K, XU M, et al. An adaptive threshold method for image denoising based on wavelet domain [C].Sixth International Symposium on Multispectral Image Processing and Pattern Recognition, International Society for Optics and Photonics, 2009: 74954M-74954M-7.[14]CARASSO A S. Direct blind deconvolution [J]. SIAM Journal on Applied Mathematics, 2001, 61(6): 1980-2007.[15]OLIVEIRA J P, FIGUEIREDO M A T, BIOUCAS-DIAS J M. Blind Estimation of Motion Blur Parameters for Image Deconvolution[M].Berlin:Springer Berlin Heidelberg, 2007: 604-611.[16]KRAHMER F, LIN Y, MCADOO B, et al.. Blind image deconvolution: Motion blur estimation [J]. IMA Preprints Series, 2006: 2133-5.[17]MOGHADDAM M E, JAMZAD M. Motion blur identification in noisy images using fuzzy sets [C].Proceedings of the Fifth IEEE International Symposium on Signal Processing and Information Technology, 2005: 862-866.[18]RICHARDSON W H. Bayesian-based iterative method of image restoration [J]. Journal of the Optical Society of America, 1972, 62(1): 55-59.[19]MIGNOTTE M. A segmentation-based regularization term for image deconvolution[J].IEEE Transactions on Image Processing, 2006, 15(7): 1973-1984..[20]SROUBEK F, MILANFAR P. Robust multichannel blind deconvolution via fast alternating minimization [J]. IEEE Transactions on Image Processing, 2012, 21(4): 1687-1700.[21]石明珠,许廷发,张坤.运动成像混合模糊的全变分图像复原[J].光学精密工程,2011, 19(8): 1973-1981. SHI M ZH, XU T F, ZHANG K. Total variation image restoration for mixed blur in moving image[J].Opt. Precision Eng., 2011, 19(8): 1973-1981.(in Chinese)[22]KRISHNAN D, FERGUS R. Fast image deconvolution using hyper-Laplacian priors [J]. Advances in Neural Information Processing Systems, 2009, 22: 1-9.[23]MARTIN D, FOWLKES C, TAL D, et al.. A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics[C].Proceedings Eighth IEEE International Conference on Computer Vision,ICCV 2001, 2: 416-423.

浏览量

252

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于跨尺度字典学习的图像盲解卷积算法

基于相位相关技术的MEMS旋转角度的高分辨力测量方法

相关系数分析在模糊图像参数识别中的应用