Real and Synthetic Underwater Image Enhancement by Colour Compensation with Complementary Colours

Underwater images undergo many light effects which degrade the quality of the image. Underwater images suffer from scattering and attenuation of light in underwater environment. Attenuation of light signal in underwater environment vary from colour to colour which complicates the process of identifying light attenuation and also enhancement or restoration. The combined effect of light attenuation and scattering will pose a difficult situation for capturing underwater images. Most of the underwater image enhancement methods depend on depth estimation and strive to calculate depth of individual pixel. Illumination of depth in not only underwater but in any image is rather difficult and subsequent processes are mostly inaccurate. In the current work, a scheme based on white balancing is proposed. First, underwater image is compensated for colour cast and two images are generated from the original image by sharpening and Gamma correction in spatial domain. In wavelet domain, these two images are fused to generate high quality enhanced image. Simulation results along with enhancement metrics prove that the proposed scheme performs well on both real underwater images and synthetic underwater images.