Enhancement of underwater images based on color compensation and color line model

Xiubo Liang, Zhigang Zhao, Zhendong Liu, Haifeng Yu, Xiaomin Yao, Xinzhe YaoJournal of Applied Optics 2025 · DOI

Abstract

针对水下光衰减和悬浮粒子导致的水下图像蓝绿色偏和雾化问题,提出了一种基于颜色补偿和颜色线模型的水下图像增强算法。首先,通过改进颜色补偿算法对图像的衰减通道进行判断和补偿,以降低图像色偏程度并提高颜色线的拟合效果。其次,结合信息量评价指标和迭代阈值策略实现图像背景分离,以避免前景明亮物体的影响,从而提高背景光估计的准确性。然后,利用颜色线与背景光矢量构建了一种能够准确估计透射率的加速凸优化算法,以提高图像的去雾效果。最后,通过对颜色通道进行归一化处理得到没有色偏的无雾图像。增强后图像的平均峰值信噪比(PSNR)、水下彩色图像质量评价(UCIQE)、水下图像质量评价(UIQM)和彩色投射因子<italic>K</italic>分别达到25.36 dB、0.64、5.02和0.02。实验结果显示,相比其他算法,本文算法能更有效地解决水下图像的颜色失真和雾化现象。
BibTeX
@article{liang2025enhancement,
  author = {Xiubo Liang and Zhigang ZHAO and Zhendong Liu and Haifeng Yu and Xiaomin Yao and Xinzhe YAO},
  title = {{Enhancement of underwater images based on color compensation and color line model}},
  journal = {Journal of Applied Optics},
  year = {2025},
  doi = {10.5768/jao202546.0502003},
}

Loading related content...

Figure