• Optics and Precision Engineering
  • Vol. 30, Issue 17, 2133 (2022)
Zhenyu HU1, Qi CHEN1,2,*, and Daqi ZHU1,2
Author Affiliations
  • 1Shanghai Engineering Research Center of Intelligent Maritime Search & Rescue and Underwater Vehicles, Shanghai Maritime University, Shanghai20306, China
  • 2School of Mechanical Engineering, University of Shanghai for Science and Technology, Shanghai00093, China
  • show less
    DOI: 10.37188/OPE.20223017.2133 Cite this Article
    Zhenyu HU, Qi CHEN, Daqi ZHU. Underwater image enhancement based on color balance and multi-scale fusion[J]. Optics and Precision Engineering, 2022, 30(17): 2133 Copy Citation Text show less
    Flow chart of underwater image enhancement algorithm based on color balance and multiscale fusion
    Fig. 1. Flow chart of underwater image enhancement algorithm based on color balance and multiscale fusion
    Schematic diagram of light absorption by water
    Fig. 2. Schematic diagram of light absorption by water
    Original and color balanced images
    Fig. 3. Original and color balanced images
    Principle of CLAHE algorithm
    Fig. 4. Principle of CLAHE algorithm
    Original image and histograms of its color channels
    Fig. 5. Original image and histograms of its color channels
    CLAHE processed image and histograms of its color channels
    Fig. 6. CLAHE processed image and histograms of its color channels
    Weight maps of two images
    Fig. 7. Weight maps of two images
    Image fusion process
    Fig. 8. Image fusion process
    Color restoration effects
    Fig. 9. Color restoration effects
    Original images and images processed by different algorithms
    Fig. 10. Original images and images processed by different algorithms
    Results of ablation experiment
    Fig. 11. Results of ablation experiment
    算 法运行时间/s
    UDCP77.435 5
    IDCPAGW10.323 1
    Multi-scale Retinex9.367 6
    CLAHE1.820 3
    本文算法20.828 5
    Table 1. Running time of different algorithms
    图 像原始值UDCPIDCPAGWMulti-scale RetinexCLAHE本文算法
    image17.172 16.608 26.234 87.326 67.453 57.796 6
    image27.485 07.424 56.729 77.109 97.731 07.875 6
    image37.063 06.010 56.344 77.057 17.474 47.731 1
    image47.132 36.759 36.584 17.202 57.545 27.808 1
    image57.261 76.885 46.657 07.181 77.594 27.801 9
    image67.202 75.678 76.888 67.145 37.376 77.747 8
    image77.377 46.183 26.853 47.173 77.518 17.818 4
    image85.656 45.686 24.220 56.667 66.894 67.712 9
    Table 2. Entropy values of images
    图 像原始值UDCPIDCPAGWMulti-scale RetinexCLAHE本文算法
    image11.048 14.950 53.364 84.274 84.291 75.221 5
    image2-0.111 22.079 52.474 80.575 03.032 53.893 6
    image30.409 23.469 03.720 14.699 32.372 94.762 7
    image42.833 34.571 54.713 25.281 64.783 35.659 6
    image50.087 00.385 43.979 83.628 22.571 64.756 2
    image62.042 84.312 73.792 42.873 83.356 44.592 4
    image71.701 13.147 04.046 52.568 43.687 44.915 1
    image81.075 94.461 92.478 93.233 91.867 04.365 3
    Table 3. UIQM values of images
    图 像原始值UDCPIDCPAGWMulti-scale RetinexCLAHE本文算法
    image10.356 20.553 80.368 20.507 00.520 60.597 3
    image20.433 80.481 90.436 00.440 40.526 90.567 5
    image30.324 10.573 70.452 10.488 00.474 40.561 9
    image40.398 50.584 80.489 90.478 30.540 50.590 7
    image50.392 80.440 90.490 80.524 50.503 50.657 3
    image60.484 60.476 80.548 50.536 90.550 00.635 0
    image70.435 80.531 50.502 80.456 90.537 00.608 2
    image80.292 70.398 00.321 00.421 10.363 40.623 7
    Table 4. UCIQE values of images
    方 法信息熵UIQMUCIQE
    原始图像6.776 01.437 40.396 5
    只用CLAHE7.583 32.597 10.487 3
    只用颜色平衡6.863 03.570 10.593 8
    本文算法7.797 54.441 20.599 0
    Table 5. Objective evaluation indexes of ablation experiment
    Zhenyu HU, Qi CHEN, Daqi ZHU. Underwater image enhancement based on color balance and multi-scale fusion[J]. Optics and Precision Engineering, 2022, 30(17): 2133
    Download Citation