Optics and Precision Engineering, Volume. 33, Issue 5, 802(2025)

Adaptive color compensation and multi-scale fusion for underwater image enhancement

Cuiyin LIU1,*... Yuanshuai CHEN1, Tengyun JING1 and Feng WANG2 |Show fewer author(s)
Author Affiliations
  • 1School of Information Engineering and Automation, Kunming University of Science and Technology, Kunming650500, China
  • 2School of Physics and Material Sciences, Guangzhou University, Guangzhou510006, China
  • show less

    To address the challenges of color distortion, uneven lighting, low contrast, and blurred details in underwater images, an adaptive color compensation and multi-scale fusion enhancement algorithm is proposed. Initially, adaptive color compensation and white light balance algorithms are employed to correct the image colors. Subsequently, adaptive gamma correction is applied to the color-corrected images to achieve uniform illumination. Thereafter, the uniformly lit images undergo processing through adaptive Rayleigh histogram stretching and adaptive unsharp masking to enhance contrast and detail. Finally, non-downsampling contour wave fusion is utilized to further refine the quality of the images with improved contrast and enhanced details. Experimental results demonstrate that the proposed algorithm effectively corrects the colors of underwater images and significantly enhances their contrast and clarity, with average evaluation metrics PSNR, SSIM, IE, and UCIQE showing increases of 43.52%, 58.65%, 12.5%, and 32.13%, respectively, compared to the original images. This algorithm is shown to substantially improve underwater image quality.

    Keywords
    Tools

    Get Citation

    Copy Citation Text

    Cuiyin LIU, Yuanshuai CHEN, Tengyun JING, Feng WANG. Adaptive color compensation and multi-scale fusion for underwater image enhancement[J]. Optics and Precision Engineering, 2025, 33(5): 802

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: Oct. 25, 2024

    Accepted: --

    Published Online: May. 20, 2025

    The Author Email: LIU Cuiyin (liucuiyin@kust.edu)

    DOI:10.37188/OPE.20253305.0802

    Topics