• Laser & Optoelectronics Progress
  • Vol. 57, Issue 12, 121102 (2020)
Dongmei Huang1、2, Yonglan Li1, Minghua Zhang1、*, and Wei Song1
Author Affiliations
  • 1College of Information Technology, Shanghai Ocean University, Shanghai 201306, China
  • 2Shanghai University of Electric Power, Shanghai 200090, China
  • show less
    DOI: 10.3788/LOP57.121102 Cite this Article Set citation alerts
    Dongmei Huang, Yonglan Li, Minghua Zhang, Wei Song. Hyperspectral Image Denoising By Combining Ground Object Features with Low-Rank Characteristics[J]. Laser & Optoelectronics Progress, 2020, 57(12): 121102 Copy Citation Text show less
    HSI low rank feature. (a) Local block; (b) object block
    Fig. 1. HSI low rank feature. (a) Local block; (b) object block
    GTLR noise reduction process. (a) Using spatial-spectral low rank of object block; (b) using global image spectral low rank
    Fig. 2. GTLR noise reduction process. (a) Using spatial-spectral low rank of object block; (b) using global image spectral low rank
    Comparison of PSNR values of noise reduction results
    Fig. 3. Comparison of PSNR values of noise reduction results
    Comparison of SSIM values of noise reduction results
    Fig. 4. Comparison of SSIM values of noise reduction results
    Comparison of FSIM values of noise reduction results
    Fig. 5. Comparison of FSIM values of noise reduction results
    Band 1 noise reduction results of Indian Pines image. (a) Original image; (b) LRMR; (c) NAILRMA; (d) proposed method
    Fig. 6. Band 1 noise reduction results of Indian Pines image. (a) Original image; (b) LRMR; (c) NAILRMA; (d) proposed method
    Band 2 noise reduction results of Indian Pines image. (a) Original image; (b) LRMR; (c) NAILRMA; (d) proposed method
    Fig. 7. Band 2 noise reduction results of Indian Pines image. (a) Original image; (b) LRMR; (c) NAILRMA; (d) proposed method
    Band 103 noise reduction results of Indian Pines image. (a) Original image; (b) LRMR; (c) NAILRMA; (d) proposed method
    Fig. 8. Band 103 noise reduction results of Indian Pines image. (a) Original image; (b) LRMR; (c) NAILRMA; (d) proposed method
    IndexLRMRNAILRMAR-GTLRC-GTLR
    SNR /dB22.832522.054721.357521.5710
    MPSNR /dB31.230933.537133.854133.9788
    MSSIM0.84280.88060.92050.9215
    MFSIM0.91440.93370.92150.9487
    Table 1. Noise reduction results for simulated noise case 1
    IndexLRMRNAILRMAR-GTLRC-GTLR
    SNR /dB24.242527.565732.245032.6148
    MPSNR /dB41.217939.598843.917444.2105
    MSSIM0.95360.96430.99560.9958
    MFSIM0.97020.97720.99490.9951
    Table 2. Noise reduction results for simulated noise case 2
    IndexLRMRNAILRMAR-GTLRC-GTLR
    SNR /dB21.804821.053727.394627.7530
    MPSNR /dB39.202638.695643.052243.3406
    MSSIM0.93520.95810.98880.9889
    MFSIM0.96270.97280.99260.9929
    Table 3. Noise reduction results for simulated noise case 3
    Dongmei Huang, Yonglan Li, Minghua Zhang, Wei Song. Hyperspectral Image Denoising By Combining Ground Object Features with Low-Rank Characteristics[J]. Laser & Optoelectronics Progress, 2020, 57(12): 121102
    Download Citation