留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于贝叶斯分层模型的MCMC方法在闪光图像重建中的应用

王忠淼 刘军 景越峰 刘进 管永红

王忠淼, 刘军, 景越峰, 等. 基于贝叶斯分层模型的MCMC方法在闪光图像重建中的应用[J]. 强激光与粒子束, 2018, 30: 114004. doi: 10.11884/HPLPB201830.180123
引用本文: 王忠淼, 刘军, 景越峰, 等. 基于贝叶斯分层模型的MCMC方法在闪光图像重建中的应用[J]. 强激光与粒子束, 2018, 30: 114004. doi: 10.11884/HPLPB201830.180123
Wang Zhongmiao, Liu Jun, Jing Yuefeng, et al. Applications of MCMC method based on Bayesian hierarchical model in flash radiography reconstruction[J]. High Power Laser and Particle Beams, 2018, 30: 114004. doi: 10.11884/HPLPB201830.180123
Citation: Wang Zhongmiao, Liu Jun, Jing Yuefeng, et al. Applications of MCMC method based on Bayesian hierarchical model in flash radiography reconstruction[J]. High Power Laser and Particle Beams, 2018, 30: 114004. doi: 10.11884/HPLPB201830.180123

基于贝叶斯分层模型的MCMC方法在闪光图像重建中的应用

doi: 10.11884/HPLPB201830.180123
基金项目: 

国家自然科学基金青年科学基金项目 11704356

详细信息
    作者简介:

    王忠淼(1994-), 男,硕士,从事闪光照相图像处理研究:neolus@qq.com

    通讯作者:

    刘军(1965-), 男,博士,研究员,从事辐射成像理论研究:liujun_ifp@126.com

  • 中图分类号: TP391

Applications of MCMC method based on Bayesian hierarchical model in flash radiography reconstruction

  • 摘要: 针对闪光图像得到的光程数据,采用贝叶斯分层模型建立了后验概率模型,运用Gibbs抽样动态构造马尔可夫链;进而获得了关于线吸收系数的统计结果及其不确定度,并与约束共轭梯度(CCG)方法进行对比分析。数值实验结果表明,马尔可夫链蒙特卡罗(MCMC)方法对理想光程图像的重建结果与真值近似完全一致;在含模糊和噪声时,重建结果与CCG方法相当;当含模糊且噪声干扰较大时,MCMC方法的重建结果要略优于CCG;更重要的是MCMC方法能够给出重建结果的不确定度。
  • 图  1  光程图像

    Figure  1.  Optical path length(OPL) images

    图  2  理想光程的重建结果

    Figure  2.  Reconstructed images of ideal OPL

    图  3  理想光程的重建结果剖面线

    Figure  3.  Section lines of reconstructions of ideal OPL

    图  4  信噪比32.3时的重建结果

    Figure  4.  Reconstructions of OPL with SNR 32.3

    图  5  信噪比为46.7时,MCMC重建中心剖面线及可信区间

    Figure  5.  Section line of reconstruction of OPL with SNR 46.7 by MCMC and its credible interval

    表  1  重建数据均方根误差比较

    Table  1.   Root mean square errors by different method

    RSN σ/%
    CCG1 CCG2 MCMC
    9.50 11.86 11.65 11.20
    12.6 10.73 10.48 10.12
    32.3 8.22 7.32 7.69
    46.7 7.93 7.00 7.39
    0 0.79
    下载: 导出CSV
  • [1] Wong A, Mishra A, Zhang W, et al. Stochastic image denoising based on Markov-chain Monte Carlo sampling[J]. Signal Processing, 2011, 91(8): 2112-2120. doi: 10.1016/j.sigpro.2011.03.021
    [2] Fadili M J, Starck J L. Sparse representation-based image deconvolution by iterative thresholding[C]//Proceedings of Astronomical Data Analysis Ada. 2006.
    [3] Cho M, Shin Y M, Lee K M. Co-recognition of image pairs by data-driven monte carlo image exploration[C]//European Conference on Computer Vision, 2009: 144-157.
    [4] Tian J, Ma K K. A MCMC approach for Bayesian super-resolution image reconstruction[C]//IEEE International Conference on Image Processing. 2005: I-45-8.
    [5] Watson S A, Balzer S, Gossein C. Density measurement errors at DARHT quantifying a decade of progress[R]. Defense Science Quarterly, 2009.
    [6] Gelman A, Carlin J B, Stern H S, et al. Bayesian data analysis[M]. 3rd Ed. Boca Raton: C Press, 2012.
    [7] Bardsley J M. MCMC-based image reconstruction with uncertainty quantification[J]. Siam Journal on Scientific Computing, 2012, 34(3): A1316-A1332. doi: 10.1137/11085760X
    [8] Howard M, Luttman A, Fowler M. Sampling-based uncertainty quantification in deconvolution of X-ray radiographs[J]. Journal of Computational & Applied Mathematics, 2014, 270(270): 43-51.
    [9] Bishop C M, Nasrabadi N M. Pattern recognition and machine learning[M]. Springer, 2006: 461-462.
    [10] Rue H, Held L. Gaussian Markov random fields: Theory and applications[J]. Technometrics, 2005, 48(1): 146-147.
    [11] Howard M, Fowler M, Luttman A, et al. Bayesian Abel inversion in quantitative X-ray radiography[J]. Siam Journal on Scientific Computing, 2016, 38(3): B396-B413. doi: 10.1137/15M1018721
    [12] 景越峰, 管永红, 张小琳. 基于约束优化的闪光照相图像重建算法[J]. 强激光与粒子束, 2016, 28: 094002. https://www.cnki.com.cn/Article/CJFDTOTAL-QJGY201609021.htm

    Jing Yuefeng, Guan Yonghong, Zhang Xiaolin. Constrained optimization reconstruction for flash radiographic image. High Power Laser and Particle Beams, 2016, 28: 094002 https://www.cnki.com.cn/Article/CJFDTOTAL-QJGY201609021.htm
    [13] 景越峰, 刘军, 管永红. 改进的约束共轭梯度闪光照相图像重建算法[J]. 强激光与粒子束, 2011, 23(8): 2201-2204. https://www.cnki.com.cn/Article/CJFDTOTAL-QJGY201108048.htm

    Jing Yuefeng, Liu Jun, Guan Yonghong. Improved constrained conjugate gradient reconstruction algorithm for flash radiographic image. High Power Laser and Particle Beams, 2011, 23(8): 2201-2204 https://www.cnki.com.cn/Article/CJFDTOTAL-QJGY201108048.htm
    [14] Odyniec M, Luttman A, Howardb M. Maximum likelihood estimation and uncertainty quantification for signals with Poisson-Gaussian mixed noise[R]. Site-Directed Research and Development, 2015: 211.
  • 加载中
图(5) / 表(1)
计量
  • 文章访问数:  1131
  • HTML全文浏览量:  281
  • PDF下载量:  90
  • 被引次数: 0
出版历程
  • 收稿日期:  2018-04-25
  • 修回日期:  2018-07-25
  • 刊出日期:  2018-11-15

目录

    /

    返回文章
    返回