Adaptive Underwater Image Enhancement via Color Channel Compensation Based on Optical Restoration and Fusion.

As human beings continue their way to find more and more resources beneath water and ocean, it becomes more urgent to have a very clear and detailed underwater image for us to explore the world unseen in the water. However, with the light propagating into the water, it is absorbed and scattered along the way which makes the underwater image unclear, hazy, detail-lost and color-shifted. Obviously, it is not the image we wish for. In the paper, the proposed method aims to enhance underwater image adaptively via color channel compensation based on optical image model. In the beginning the underwater image is restored only in green channel aiming to reduce the haze effect, then an adaptive color channel compensation is applied to correct the shifted color, lastly a multi-scale fusion is executed to show more image details after a white balance operation. Going for massive experiments, the proposed adaptive method fits in versatile scenes adaptively of greenish, bluish and turbid water body producing eye-friendly haze cover removed, color shift corrected, detail enhanced clear result image. Particularly the proposed method highly reduces the reddish effect after execution compared to many other state of art underwater image enhancement algorithm, while quantitatively the proposed method gives a better score too by underwater image quality measure (UIQM) and underwater color image quality evaluation (UCIQE).