• Laser & Optoelectronics Progress
  • Vol. 60, Issue 4, 0401003 (2023)
Guodong Liu1, Lihui Feng1、*, Jihua Lu2、**, and Jianmin Cui1
Author Affiliations
  • 1School of Optics and Photonics, Beijing Institute of Technology, Beijing 100081, China
  • 2School of Integrated Circuits and Electronics, Beijing Institute of Technology, Beijing 100081, China
  • show less
    DOI: 10.3788/LOP220651 Cite this Article Set citation alerts
    Guodong Liu, Lihui Feng, Jihua Lu, Jianmin Cui. Underwater Image Restoration Based on Classification and Dark Channel Prior with Minimum Convolutional Area[J]. Laser & Optoelectronics Progress, 2023, 60(4): 0401003 Copy Citation Text show less
    Schematic diagram of underwater formation model
    Fig. 1. Schematic diagram of underwater formation model
    Flowchart of HSV-CIELAB classification equalization and minimum convolution area DCP
    Fig. 2. Flowchart of HSV-CIELAB classification equalization and minimum convolution area DCP
    CIELAB color equalization results. (a) RGB channel grayscale of raw image; (b) RGB channel grayscale of balanced image
    Fig. 3. CIELAB color equalization results. (a) RGB channel grayscale of raw image; (b) RGB channel grayscale of balanced image
    Comparison of LAB histograms of images. (a) Channel L of raw image; (b) channel a of raw image; (c) channel b of raw image; (d) channel L of balanced image; (e) channel a of balanced image; (f) channel b of balanced image
    Fig. 4. Comparison of LAB histograms of images. (a) Channel L of raw image; (b) channel a of raw image; (c) channel b of raw image; (d) channel L of balanced image; (e) channel a of balanced image; (f) channel b of balanced image
    Composite false color image
    Fig. 5. Composite false color image
    Backlight estimation. (a) Original image; (b) dark channel Dx; (c) convolution image G x
    Fig. 6. Backlight estimation. (a) Original image; (b) dark channel Dx; (c) convolution image G  x
    Example of backlight estimation with minimum convolution area DCP algorithm
    Fig. 7. Example of backlight estimation with minimum convolution area DCP algorithm
    Comparison of visual results of different algorithms. (a) (b) High-saturation distorted images; (c) (d) low-saturation distortion images; (e) (f) shallow water images
    Fig. 8. Comparison of visual results of different algorithms. (a) (b) High-saturation distorted images; (c) (d) low-saturation distortion images; (e) (f) shallow water images
    Enhancement results of different algorithms and comparison with reference images. (a) (b) (c) High-saturation distorted images; (d) (e) (f) (g) (h) low-saturation distorted images; (i) shallow water image
    Fig. 9. Enhancement results of different algorithms and comparison with reference images. (a) (b) (c) High-saturation distorted images; (d) (e) (f) (g) (h) low-saturation distorted images; (i) shallow water image
    AlgorithmParameterUDCPULAPIBLAMIPProposed algorithm
    Image(a)PSNR10.891118.894718.787715.424419.1593
    SSIM0.48920.73770.76940.69690.7298
    Image(b)PSNR12.926917.503720.730419.320321.1145
    SSIM0.67790.70330.82950.75300.8716
    Image(c)PSNR12.538416.521317.451719.175421.9297
    SSIM0.63350.69130.63560.80480.8136
    Image(d)PSNR10.336417.096219.312316.628321.8104
    SSIM0.60460.79370.84630.77320.8943
    Image(e)PSNR15.793618.863017.614817.747822.7959
    SSIM0.51970.80530.87040.85020.8880
    Image(f)PSNR15.821917.345616.453916.759118.3918
    SSIM0.73150.85100.80770.79590.8590
    Image(g)PSNR11.110515.979516.361416.142420.3412
    SSIM0.69590.83650.83510.82620.8878
    Image(h)PSNR18.040521.397320.764221.145522.7646
    SSIM0.83910.91160.92830.93380.9188
    Image(i)PSNR13.544213.619615.822713.573718.6227
    SSIM0.73190.76330.61390.75660.8931
    Table 1. PSNR and SSIM evaluation quality of different algorithms
    AlgorithmUDCPULAPIBLAMIPProposed algorithmReference
    Image(a)0.46290.53400.56220.52750.57100.5777
    Image(b)0.57510.58270.60990.61050.59340.5778
    Image(c)0.48690.55270.47440.54510.57560.5718
    Image(d)0.45920.57120.55020.51090.57460.5581
    Image(e)0.53190.57380.59120.58870.58290.6139
    Image(f)0.53860.58490.59060.55450.58370.5031
    Image(g)0.55700.58170.58640.58980.60810.6074
    Image(h)0.58180.60220.56800.58620.56850.5670
    Image(i)0.52440.54650.53150.54980.55410.5287
    Table 2. Parameters of UCIQE of different algorithms
    Guodong Liu, Lihui Feng, Jihua Lu, Jianmin Cui. Underwater Image Restoration Based on Classification and Dark Channel Prior with Minimum Convolutional Area[J]. Laser & Optoelectronics Progress, 2023, 60(4): 0401003
    Download Citation