• Laser & Optoelectronics Progress
  • Vol. 57, Issue 8, 082801 (2020)
Zanwei Yang1、2, Liangliang Zheng1、*, Yong Wu1, and Hongsong Qu1
Author Affiliations
  • 1Department of Advanced Space Technology, Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin 130033, China
  • 2University of Chinese Academy of Sciences, Beijing 100049, China
  • show less
    DOI: 10.3788/LOP57.082801 Cite this Article Set citation alerts
    Zanwei Yang, Liangliang Zheng, Yong Wu, Hongsong Qu. An Improved Moment Matching Algorithm for Non-Uniform Correction of Hyperspectral Images[J]. Laser & Optoelectronics Progress, 2020, 57(8): 082801 Copy Citation Text show less
    Flow chart of improved moment matching non-uniformity correction algorithm
    Fig. 1. Flow chart of improved moment matching non-uniformity correction algorithm
    Simulated images. (a) Original image; (b) image with stripe noise
    Fig. 2. Simulated images. (a) Original image; (b) image with stripe noise
    Destriped results of simulated image in Fig. 2(b) by different methods. (a) BW; (b) WMM; (c) DSLFRI; (d) HM; (e) proposed method
    Fig. 3. Destriped results of simulated image in Fig. 2(b) by different methods. (a) BW; (b) WMM; (c) DSLFRI; (d) HM; (e) proposed method
    Gray mean values of simulated image in column.(a)Original image;(b) image with stripe noise
    Fig. 4. Gray mean values of simulated image in column.(a)Original image;(b) image with stripe noise
    Gray mean values of destriped results in column by different methods. (a) BW; (b) WMM; (c) DSLFRI; (d) HM; (e) proposed method
    Fig. 5. Gray mean values of destriped results in column by different methods. (a) BW; (b) WMM; (c) DSLFRI; (d) HM; (e) proposed method
    Destriped results of hyperspectral image of band 25 by different methods. (a) Original image; (b)BW; (c) WMM; (d) DSLFRI; (e) HM; (f) proposed method
    Fig. 6. Destriped results of hyperspectral image of band 25 by different methods. (a) Original image; (b)BW; (c) WMM; (d) DSLFRI; (e) HM; (f) proposed method
    Destriped results of hyperspectral image of band 27 by different methods. (a) Original image; (b) BW; (c) WMM; (d) DSLFRI; (e) HM; (f) proposed method
    Fig. 7. Destriped results of hyperspectral image of band 27 by different methods. (a) Original image; (b) BW; (c) WMM; (d) DSLFRI; (e) HM; (f) proposed method
    Gray mean scale of destriped results of hyperspectral image of band 25 by different methods. (a) Original image; (b) BW; (c) WMM; (d) DSLFRI; (e) HM; (f) proposed method
    Fig. 8. Gray mean scale of destriped results of hyperspectral image of band 25 by different methods. (a) Original image; (b) BW; (c) WMM; (d) DSLFRI; (e) HM; (f) proposed method
    Gray mean values of destriped results of hyperspectral image of band 27 by different methods. (a) Original image; (b) BW; (c) WMM; (d) DSLFRI; (e) HM; (f) proposed method
    Fig. 9. Gray mean values of destriped results of hyperspectral image of band 27 by different methods. (a) Original image; (b) BW; (c) WMM; (d) DSLFRI; (e) HM; (f) proposed method
    MethodBWWMMDSLFRIHMProposed method
    MSE21.05507.83568.053317.77131.8726
    PSNR /dB34.897239.190139.071135.633645.4064
    SSIM0.57440.96310.96490.73360.9903
    Table 1. Comparison of assessment criteria for simulated image in Fig. 2(b)
    Destripe noiseIndexBWWMMDSLFRIHMProposed method
    Random noise in [10,40]MSE20.13566.19877.072735.20711.9322
    PSNR35.091240.207839.634932.664545.2703
    SSIM0.58270.96720.96580.73690.9873
    Random noise in [20,40]MSE21.05507.83568.053317.77131.8726
    PSNR34.897239.190139.071135.633645.4064
    SSIM0.57440.96310.96490.73360.9903
    Random noise in [30,40]MSE22.744010.915611.012629.53071.7772
    PSNR34.562137.750337.711933.428145.6335
    SSIM0.55890.96120.96360.65540.9890
    Periodic noise in [10,40]MSE19.75784.60476.038913.66091.7803
    PSNR35.173441.496340.321236.776045.6259
    SSIM0.57860.97000.96530.79940.9918
    Periodic noise in [20,40]MSE20.11465.12816.410414.78421.7127
    PSNR35.095741.031340.061936.432845.7941
    SSIM0.56670.96810.96540.76670.9921
    Periodic noise in [30,40]MSE20.23255.30426.579115.39841.6796
    PSNR35.070340.884639.949136.256145.8788
    SSIM0.56050.96790.96500.74920.9924
    Table 2. Comparison of objective indexes for simulated images with different noises
    MethodOriginal imageBWWMMDSLFRIHMProposed method
    ICV6.98146.86347.06897.16787.36717.6435
    DEC160.0580163.9211156.0881151.8557141.0393136.6519
    RM2.77920.80472.60962.60592.65772.4399
    Table 3. Comparison of objective indexes for hyperspectral image of band 25
    MethodOriginal imageBWWMMDSLFRIHMProposed method
    ICV8.77648.47868.93388.58838.65299.7724
    DEC116.2303123.5045112.125897.3851100.054494.8759
    RM2.92380.93462.78182.76763.24592.5334
    Table 4. Comparison of objective indexes for hyperspectral image of band 27
    Zanwei Yang, Liangliang Zheng, Yong Wu, Hongsong Qu. An Improved Moment Matching Algorithm for Non-Uniform Correction of Hyperspectral Images[J]. Laser & Optoelectronics Progress, 2020, 57(8): 082801
    Download Citation