基于自适应阈值的干涉高光谱图像稀疏重建

温佳, 刘明威, 崔军, 等. 基于自适应阈值的干涉高光谱图像稀疏重建[J]. 光电工程, 2019, 46(6): 180443. doi: 10.12086/oee.2019.180443
引用本文: 温佳, 刘明威, 崔军, 等. 基于自适应阈值的干涉高光谱图像稀疏重建[J]. 光电工程, 2019, 46(6): 180443. doi: 10.12086/oee.2019.180443
Wen Jia, Liu Mingwei, Cui Jun, et al. Sparse reconstruction of interferometric hyperspectral image based on adaptive threshold[J]. Opto-Electronic Engineering, 2019, 46(6): 180443. doi: 10.12086/oee.2019.180443
Citation: Wen Jia, Liu Mingwei, Cui Jun, et al. Sparse reconstruction of interferometric hyperspectral image based on adaptive threshold[J]. Opto-Electronic Engineering, 2019, 46(6): 180443. doi: 10.12086/oee.2019.180443

基于自适应阈值的干涉高光谱图像稀疏重建

  • 基金项目:
    天津市自然科学基金项目(17JCQNJC01400);国家自然科学基金资助项目(61401439,61601323)
详细信息
    作者简介:
    通讯作者: 刘明威(1992-),男,硕士研究生,主要从事压缩感知,图像处理的研究。E-mail:mr.liumw@qq.com
  • 中图分类号: O433.4

Sparse reconstruction of interferometric hyperspectral image based on adaptive threshold

  • Fund Project: Supported by Natural Science Foundation of Tianjing (17JCQNJC01400) and National Natural Science Foundation of China (61401439, 61601323)
More Information
  • 干涉高光谱图像是一类特殊的图像源,其海量数据导致很难在有限带宽信道上传输。传统的方法是对数据进行压缩,然后进行编码传输。但是压缩后的数据还是很大,给数据的传输和存储带来很大困难,而压缩感知技术可以很好地解决该类图像在传输时的问题。本文在压缩感知原有算法的基础上提出了更适用于干涉高光谱图像的基于自适应阈值的正交匹配追踪算法(ATROMP),该算法首先采用分块处理,然后挑选出干涉条纹块。由于竖直干涉条纹具有较强的单方向特性,水平全变分值较大。因此本文根据水平全变分值提取出图像中的干涉条纹,进行自适应采样。然后采用一个自适应阈值来代替正则正交匹配追踪(ROMP)算法中的二次选取,采用自适应阈值不仅可以保障每次选取的原子的相关性足够高,而且每次可以适当地选取多个原子保证足够的循环次数,避免了后续匹配度更高原子的遗漏。相比于传统ROMP算法,大量实验数据表明本文方法稀疏重建的精度可以得到明显的提高。

  • Overview: The interference hyperspectral image data is a three-dimensional image data generated by satellite scanning by a Large Aperture Static Imaging Spectrometer (LASIS) based on the principle of push-scan Fourier transform imaging. The resolution is extremely high, and its massive amount Data poses a certain degree of difficulty for data storage and transmission over limited bandwidth channels. Therefore, it is imperative to design an efficient transmission method suitable for interfering hyperspectral data for its data characteristics. Compressed sensing, as a new theoretical framework, provides new research ideas for signal description and processing. Unlike the existing sampling theorem, the theory samples the signal using a rate much smaller than the Nyquist sampling law, and then reconstructs the original signal with high probability from these small observations. This efficient sampling method greatly reduces the sampling rate, so it has great application prospects in many research fields. Based on the traditional compressed sensing reconstruction algorithm, this paper proposes a reconstruction method for interference hyperspectral image. The interference hyperspectral image is a three-dimensional image with multi-dimensional correlation, and its interference fringes contain abundant spectral information. However, when using traditional ROMP algorithm to reconstruct the image, the absolute value of the inner product of the measurement matrix and the residual needs to be calculated. As the interference hyperspectral image has interference fringes with large fluctuations in the fixed amplitude, the variance of the calculation result of the inner product is large, which will result in too many atoms to be selected in the secondary selection according to the regularization standard in each iteration. The atomic number with higher matching degree in the subsequent phase is not selected, resulting in support. The proportion of atoms with high degree of central matching is low. This will seriously affect the reconstruction quality of interference hyperspectral, especially the interference fringe. To solve the above problems, in this paper we propose an adaptive threshold regularized orthogonal matching pursuit algorithm (ATROMP). The algorithm first uses block processing and then selects the interference fringes. Because the vertical interference fringes have strong unidirectional characteristics, the interference fringes in the images are extracted from the horizontal total variation values for adaptive sampling. Then an adaptive threshold is used in this paper to replace the quadratic selection in the ROMP algorithm. Using an adaptive threshold can not only ensure that the atomicity of each selected atom is sufficiently high, but also that multiple atoms can be properly selected each time to ensure sufficient number of cycles, to avoid the follow-up higher degree of atom missing. Compared with the traditional ROMP algorithm, a large amount of experimental datas show that the sparse reconstruction accuracy of the method can be significantly improved.

  • 加载中
  • 图 1  等效光路图

    Figure 1.  Schematic diagram of LASIS

    图 2  LASIS干涉高光谱图像序列

    Figure 2.  LASIS interference hyperspectral image sequences

    图 3  本文提出的最优阈值系数算法流程图

    Figure 3.  The flow chart of the optimal threshold coefficient algorithm proposed in this paper

    图 4  ATROMP算法流程图

    Figure 4.  The flow chart of the ATROMP algorithm

    图 5  原始图像。(a) Lasis01;(b) Lasis02

    Figure 5.  Original image. (a) Lasis01; (b) Lasis02

    图 6  Lasis01的CS重构图像。(a)采样率0.35 ROMP算法;(b)采样率0.5 ROMP算法;(c)采样率0.35 ATROMP算法;(d)采样率0.5 ATROMP算法

    Figure 6.  CS reconstruction image of Lasis01. (a) Sampling rate 0.35 in ROMP algorithm; (b) Sampling rate 0.5 in ROMP algorithm; (c) Sampling rate 0.35 in ATROMP algorithm; (d) Sampling rate 0.5 in ATROMP algorithm

    图 7  Lasis02的CS重构图像。(a)采样率0.35 ROMP算法;(b)采样率0.5 ROMP算法;(c)采样率0.35 ATROMP算法;(d)采样率0.5 ATROMP算法

    Figure 7.  CS reconstruction image of Lasis02. (a) Sampling rate 0.35 in ROMP algorithm; (b) Sampling rate 0.5 in ROMP algorithm; (c) Sampling rate 0.35 in ATROMP algorithm; (d) Sampling rate 0.5 in ATROMP algorithm

    表 1  均匀采样时具体采样数据

    Table 1.  Specific sampling data for uniform sampling

    平均采样率 0.3 0.35 0.4 0.5
    干涉条纹块 150/256 160/256 170/256 200/256
    非干涉条纹块 72/256 85/256 98/256 123/256
    下载: 导出CSV

    表 2  采样率0.35和0.5时三幅图像不同算法下信噪比和结构相似度(SSIM)对比表

    Table 2.  Comparison of SNR and reconstruction time of three images under different algorithms at sampling rate 0.35 and 0.5

    采样率0.35 采样率0.5
    ROMP算法 ATROMP算法 ROMP算法 ATROMP算法
    信噪比/dB SSIM 信噪比/dB SSIM 信噪比/dB SSIM 信噪比/dB SSIM
    Lasis01 26.4556 0.8293 31.0979 0.9232 29.5770 0.9048 32.9129 0.9476
    Lasis02 26.7208 0.8321 31.4439 0.9291 29.7985 0.8918 31.7564 0.9354
    Lasis03 26.9426 0.8603 31.2438 0.9265 29.8668 0.9077 31.5194 0.9318
    下载: 导出CSV
  • [1]

    童庆禧, 张兵, 张立福.中国高光谱遥感的前沿进展[J].遥感学报, 2016, 20(5): 689-707. doi: 10.11834/jrs.20166264

    Tong Q X, Zhang B, Zhang L F. Current progress of hyperspectral remote sensing in China[J]. Journal of Remote Sensing, 2016, 20(5): 689-707. doi: 10.11834/jrs.20166264

    [2]

    汤毅, 万建伟, 粘永健.高光谱图像的分布式近无损压缩[J].光学学报, 2015, 35(3): 0310001. 10.3788/AOS201535.0310001

    Tang Y, Wan J W, Nian Y J. Distributed near lossless compression of hyperspectral images[J]. Acta Optica Sinica, 2015, 35(3): 0310001. 10.3788/AOS201535.0310001

    [3]

    王忠良, 冯燕, 肖华, 等.高光谱图像的分布式压缩感知成像与重构[J].光学精密工程, 2015, 23(4): 1131-1137. doi: 10.3788/OPE.20152304.1131

    Wang Z L, Feng Y, Xiao H, et al. Distributed compressive sensing imaging and reconstruction of hyperspectral imagery[J]. Optics and Precision Engineering, 2015, 23(4): 1131-1137. doi: 10.3788/OPE.20152304.1131

    [4]

    杜述松, 王咏梅, 王英鉴.空间应用干涉成像光谱仪的研究[J].光学仪器, 2008, 30(3): 77-82. doi: 10.3969/j.issn.1005-5630.2008.03.019

    Du S S, Wang Y M, Wang Y J. The research of the space applied interference imaging spectrometer[J]. Optical Instruments, 2008, 30(3): 77-82. doi: 10.3969/j.issn.1005-5630.2008.03.019

    [5]

    温佳, 赵军锁, 王彩玲, 等.基于改进MCA的干涉高光谱图像分解[J].光谱学与光谱分析, 2016, 36(1): 254-258. 10.3964/j.issn.1000-0593(2016)01-0254-05

    Wen J, Zhao J S, Wang C L, et al. Decomposition of interference hyperspectral images using improved morphological component analysis[J]. Spectroscopy and Spectral Analysis, 2016, 36(1): 254-258. 10.3964/j.issn.1000-0593(2016)01-0254-05

    [6]

    吴冬梅, 王军, 张海宁.基于谱间DPCM和整数小波变换的超光谱图像无损压缩[J].光子学报, 2008, 37(1): 156-159. http://d.old.wanfangdata.com.cn/Periodical/gzxb200801034

    Wu D M, Wang J, Zhang H N. Research of hyper-spectral image lossless compressing technology based on DPCM between spectra and IWT[J]. Acta Photonica Sinica, 2008, 37(1): 156-159. http://d.old.wanfangdata.com.cn/Periodical/gzxb200801034

    [7]

    Wen J, Ma C W, Shui P L. A 3D non-linear orientation prediction wavelet transform for interference hyperspectral images compression[J]. Optics Communications, 2011, 284(7): 1770-1777. doi: 10.1016/j.optcom.2010.11.069

    [8]

    Wen J, Ma C W, Shui P L. An adaptive OPD and dislocation prediction used characteristic of interference pattern for interference hyperspectral image compression[J]. Optics Communications, 2011, 284(20): 4903-4909. doi: 10.1016/j.optcom.2011.06.061

    [9]

    Li Y, Chen L W, Kong F, et al. Functional micro-concrete 3D hybrid structures fabricated by two-photon polymerization[J]. Opto-Electronic Engineering, 2017, 44(4): 393-399. 10.3969/j.issn.1003-501X.2017.04.002

    [10]

    Donoho D L. Compressed sensing[J]. IEEE Transactions on Information Theory, 2006, 52(4): 1289-1306. doi: 10.1109/TIT.2006.871582

    [11]

    Candes E J, Romberg J, Tao T. Robust uncertainty principles: exact signal reconstruction from highly incomplete frequency information[J]. IEEE Transactions on Information Theory, 2006, 52(2): 489-509. doi: 10.1109/TIT.2005.862083

    [12]

    闫歌, 许廷发, 马旭, 等.动态测量的高光谱图像压缩感知[J].中国光学, 2018, 11(4): 550-559. doi: 10.3788/co.20181104.0550

    Yan G, Xu T F, Ma X, et al. Hyperspectral image compression sensing based on dynamic measurement[J]. Chinese Optics, 2018, 11(4): 550-559. doi: 10.3788/co.20181104.0550

    [13]

    Mun S, Fowler J E. Block compressed sensing of images using directional transforms[C]//Proceedings of 2009 IEEE International Conference on Image Processing, Cairo, Egypt, 2009: 3021-3024.https://ieeexplore.ieee.org/document/5414429

    [14]

    王荣芳, 焦李成, 刘芳, 等.利用纹理信息的图像分块自适应压缩感知[J].电子学报, 2013, 41(8): 1506-1514. doi: 10.3969/j.issn.0372-2112.2013.08.009

    Wang R F, Jiao L C, Liu F, et al. Block-based adaptive compressed sensing of image using texture information[J]. Acta Electronica Sinica, 2013, 41(8): 1506-1514. doi: 10.3969/j.issn.0372-2112.2013.08.009

    [15]

    马春晖.压缩感知重构算法研究[D].杭州: 杭州电子科技大学, 2012.

    Ma C H. Study on the compressed sensing reconstruction algorithms[D]. Hangzhou: Hangzhou Dianzi University, 2012.http://cdmd.cnki.com.cn/Article/CDMD-10336-1012323635.htm

    [16]

    贾应彪, 冯燕, 王忠良, 等.基于谱间结构相似先验的高光谱压缩感知重构[J].电子与信息学报, 2014, 36(6): 1406-1412. doi: 10.3724/SP.J.1146.2013.01132

    Jia Y B, Feng Y, Wang Z L, et al. Hyperspectral compressive sensing recovery via spectrum structure similarity[J]. Journal of Electronics & Information Technology, 2014, 36(6): 1406-1412. doi: 10.3724/SP.J.1146.2013.01132

    [17]

    Kittle D, Choi K, Wagadarikar A, et al. Multiframe image estimation for coded aperture snapshot spectral imagers[J]. Applied Optics, 2010, 49(36): 6824-6833. doi: 10.1364/AO.49.006824

    [18]

    Wu H L, Wang S. Adaptive sparsity matching pursuit algorithm for sparse reconstruction[J]. IEEE Signal Processing Letters, 2012, 19(8): 471-474. doi: 10.1109/LSP.2012.2188793

    [19]

    Zhang L, Zhang S C, Bin X L, et al. The analysis of the interference hyperspectral image features and compression[J]. Proceedings of SPIE, 2010, 7658: 765839. doi: 10.1117/12.867626

  • 加载中

(7)

(2)

计量
  • 文章访问数:  6579
  • PDF下载数:  2219
  • 施引文献:  0
出版历程
收稿日期:  2018-08-24
修回日期:  2018-12-06
刊出日期:  2019-06-01

目录

/

返回文章
返回