吉林大学学报(工学版) ›› 2015, Vol. 45 ›› Issue (4): 1360-1366.doi: 10.13229/j.cnki.jdxbgxb201504048

• • 上一篇    下一篇

结合噪声去除的极大似然图像复原

姜超1, 耿则勋1, 刘立勇2, 潘映峰3   

  1. 1.解放军信息工程大学 地理空间信息学院,郑州 450052;
    2.中国科学院 国家天文台,北京 100012;
    3. 中国人民解放军61175部队,南京 210049
  • 收稿日期:2014-03-24 出版日期:2015-07-01 发布日期:2015-07-01
  • 作者简介:姜超(1985-),男,博士研究生.研究方向:数字图像处理.E-mail:jiangchao19850429@126.com
  • 基金资助:
    “863”国家高技术研究发展计划项目(2012AA7032031D); 国家自然科学基金项目(11373043)

Maximum likelihood image restoration combined with image denoising

JIANG Chao1, GENG Ze-xun1, LIU Li-yong2, PAN Ying-feng3   

  1. 1.PLA Information Engineering University, Zhengzhou 450052, China;
    2.National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100012;
    3.Troops 61175 of PLA, Nanjing 210049, China
  • Received:2014-03-24 Online:2015-07-01 Published:2015-07-01

摘要: 基于混合噪声模型的极大似然算法不仅没有充分考虑迭代过程中的噪声影响,而且假定点扩散函数(PSF)已知且不随迭代过程变化,从而导致复原过程不稳定。在图像含噪且PSF未知的情况下,提出以去噪算法作为预处理手段,同时将PSF参数估计引入极大似然算法迭代过程并随迭代过程动态更新,最后将估计的PSF代入维纳滤波以提高复原图像的质量。实验结果证明,本文复原图像质量有明显改善,表明该算法具有较强的稳定性和抗噪声能力,是一种有效的图像复原方法。

关键词: 摄影测量与遥感技术, 图像复原, 混合噪声模型, 图像去噪, 点扩散函数, 极大似然

Abstract: In the maximum likelihood algorithm proposed by Benvenuto base on mixed noise model, the noise effect during iteration is not taken into consideration and the Point Spread Function (PSF) is assumed to be known and unchanged. This leads to the unstability of the image restoration. Under the condition of noise existence and PSF unknown, an image denoising algorithm is proposed as preprocessing, and the parameter estimation of PSF is introduced into iteration of the maximum likelihood and is dynamically updated. Finally, the Wiener filter with estimated PSF is utilized to improve the quality of the restored image. Experiment result demonstrate that the quality of the restored image is obviously improved, which proves the stability and noise resistance of the proposed algorithm.

Key words: photogrammetry and remote sensing technology, image restoration, mixed noise model, image denoising, point spread function, maximum likelihood

中图分类号: 

  • TP751
[1] Ayers C R, Dainty J C. Iterative blind deconvolution method and its application[J].Optics Letters,1988,13(7):547-549.
[2] Maccallum B C. Blind deconvolution by simulated annealing[J].Optics Communication,1990,75(2):101-105.
[3] Kundur D, Hatzinakos D. Blind image deconvolution[J].IEEE Single Processing Magazine,1996, 13(3):43-64.
[4] Benvenuto F, Camera A L. The study of an iterative method for the reconstructionof images corrupted by Poisson and Gaussian noise[J].Inverse Problems,2008,24(3):35016-35035.
[5] Rudin L I, Osher S, Fatemi E. Nonlinear total variation based noise removal algorithms[J].Physica D: Nonlinear Phenomena,1992,60:259-268.
[6] Portilla J, Strela V, Wainwright M J, et al. Image denoising using Gaussian scale mixtures in the wavelet domain[J]. IEEE Trans Image Processing,2003,12(11):1338-1351.
[7] Buades A, Coll B, Morel J M. A non local algorithm for image denoising[J].IEEE Computer Vision and Pattern Recognition,2005,2:60-65.
[8] Dabov K, Foi A, Katkovnik V, et al. Image denoising by sparse 3D transform-domain collaborative filtering[J].IEEE Trans Image Process, 2007, 16(8):2080-2095.
[9] 周宏潮, 朱炬波, 王正明. 混合泊松-高斯分布模型的参数估计[J].中国空间科学技术,2005, 25(2):1-5. Zhou Hong-chao, Zhu Ju-bo, Wang Zheng-ming. Parametric estimation of mixed Poisson-Gaussian distribution model[J].Chinese Space Science and Technology, 2005, 25(2):1-5.
[10] 陈波. 自适应光学图像复原理论与算法研究[D].郑州:解放军信息工程大学,2008:63-68. Chen Bo. The theory and algorithms of adaptive optics image restoration[D].Zhengzhou: PLA Information Engineering University,2008:63-68.
[1] 王新华, 郝建坤, 黄玮, 欧阳继红. 基于简单透镜计算成像的图像复原重建[J]. 吉林大学学报(工学版), 2017, 47(3): 965-972.
[2] 李晖晖, 滑立, 杨宁, 刘坤. 基于MSA特征和模拟退火优化的遥感图像多目标关联算法[J]. 吉林大学学报(工学版), 2015, 45(4): 1353-1359.
[3] 王京萌, 张爱武, 赵宁宁, 孟宪刚. 斜采样的倾斜角度对采样产生混叠的影响及其与分辨率的关系[J]. 吉林大学学报(工学版), 2015, 45(3): 953-960.
[4] 曹建农, 郭佳, 王蓓, 董昱威, 王平禄. 高分辨率影像中城区树冠多尺度聚类识别方法[J]. 吉林大学学报(工学版), 2014, 44(4): 1215-1224.
[5] 祁冰, 孔韦韦. 基于改进型交叉视觉皮层模型的自适应图像去噪[J]. 吉林大学学报(工学版), 2014, 44(01): 184-190.
[6] 李一兵, 付强, 张静. 水下模糊图像参数估计复原方法[J]. 吉林大学学报(工学版), 2013, 43(04): 1133-1138.
[7] 尹奎英1,刘宏伟2,金林1. 快速的Otsu双阈值SAR图像分割法[J]. 吉林大学学报(工学版), 2011, 41(6): 1760-1765.
[8] 张喜清1,项昌乐2,刘辉2. 履带车辆变速箱箱体的试验模态参数识别及验证[J]. 吉林大学学报(工学版), 2011, 41(4): 927-931.
[9] 陈玫玫, 郭树旭, 王瑶, 吴斌, 于思瑶, 邵向鑫. 压缩感知的指静脉图像去噪[J]. 吉林大学学报(工学版), 2011, 41(02): 559-0562.
[10] 王宏志,武伟,钟诚 . 基于非线性扩散与小波变换的混合图像去噪算法[J]. 吉林大学学报(工学版), 2009, 39(02): 525-0529.
[11] 崔艳秋;王珂;孙巍. 基于小波域自适应高斯混合模型的图像去噪方法[J]. 吉林大学学报(工学版), 2006, 36(06): 983-0988.
[12] 王洪民, 张魁元, 梁瑞梅. 极值分布在无失效数据下的修正极大似然估计[J]. 吉林大学学报(工学版), 2001, (1): 52-56.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!