• Laser & Optoelectronics Progress
  • Vol. 60, Issue 16, 1610006 (2023)
Haichen Wang1, Shengqi Wang1, and Xueyou Hu2,*
Author Affiliations
  • 1College of Energy Materials and Chemical Engineering, Hefei University, Hefei 230601, Anhui, China
  • 2School of Advanced Manufacturing Engineering, Hefei University, Hefei 230601, Anhui, China
  • show less
    DOI: 10.3788/LOP222268 Cite this Article Set citation alerts
    Haichen Wang, Shengqi Wang, Xueyou Hu. Optimization of Hyperspectral Image Denoising Based on Local Truncated Nuclear Norm[J]. Laser & Optoelectronics Progress, 2023, 60(16): 1610006 Copy Citation Text show less
    SSIM under different trancated values t
    Fig. 1. SSIM under different trancated values t
    PSNR under different truncated values t
    Fig. 2. PSNR under different truncated values t
    SSIM under different expected rank r
    Fig. 3. SSIM under different expected rank r
    PSNR under different expected rank r
    Fig. 4. PSNR under different expected rank r
    SSIM of the algorithm before and after improvement
    Fig. 5. SSIM of the algorithm before and after improvement
    PSNR of the algorithm before and after improvement
    Fig. 6. PSNR of the algorithm before and after improvement
    Denoising results of each denoising method in band 2 of Pavia University dataset. (a) Original image; (b) TNN-LLRGTV; (c) LLRGTV; (d) LRTDTV; (e) LRMR; (f) NAILRMA
    Fig. 7. Denoising results of each denoising method in band 2 of Pavia University dataset. (a) Original image; (b) TNN-LLRGTV; (c) LLRGTV; (d) LRTDTV; (e) LRMR; (f) NAILRMA
    Denoising results of each denoising method in band 110 of Salinas dataset. (a) Original image; (b) TNN-LLRGTV; (c) LLRGTV; (d) LRTDTV; (e) LRMR; (f) NAILRMA
    Fig. 8. Denoising results of each denoising method in band 110 of Salinas dataset. (a) Original image; (b) TNN-LLRGTV; (c) LLRGTV; (d) LRTDTV; (e) LRMR; (f) NAILRMA
    Comparison of spectral curves of different denoising algorithms
    Fig. 9. Comparison of spectral curves of different denoising algorithms
    Noise intensityParameterNAILRMALRMRLRTDTVLLRGTVTNN-LLRGTV

    G=0.04,

    S=0.10

    SSIM0.60410.66020.65120.78090.8445
    PSNR /dB22.379026.597922.714629.770230.7610

    G=0.08,

    S=0.15

    SSIM0.46270.52250.53130.67400.7706
    PSNR /dB18.671523.709618.960327.650528.7855

    G=0.12,

    S=0.20

    SSIM0.36520.41710.43990.57530.7051
    PSNR /dB16.354421.524316.565925.320826.4233
    Table 1. Denoising result for mixed noise with equal intensity
    Noise intensityParameterNAILRMALRMRLRTDTVLLRGTVTNN-LLRGTV
    G=0.04SSIM0.76720.70210.76700.86610.8836
    PSNR /dB27.299227.710827.129027.548527.4349
    G=0.08SSIM0.65780.59630.68110.71590.7993
    PSNR /dB23.493525.511523.554928.340129.3462
    G=0.12SSIM0.58230.52740.62320.65750.7554
    PSNR /dB21.273924.256821.410327.224828.2942
    S=0.10SSIM0.76480.93690.77420.95660.9569
    PSNR /dB26.993136.720627.060137.268837.3276
    S=0.15SSIM0.68740.91350.71130.95580.9566
    PSNR /dB24.231135.061924.503337.093837.2341
    S=0.20SSIM0.61840.88800.65180.95500.9561
    PSNR /dB22.124833.623122.448136.916337.1171
    Table 2. Denoising results for single-type noise
    Haichen Wang, Shengqi Wang, Xueyou Hu. Optimization of Hyperspectral Image Denoising Based on Local Truncated Nuclear Norm[J]. Laser & Optoelectronics Progress, 2023, 60(16): 1610006
    Download Citation