Underwater Image Enchancement using CNN based Learning Methods

Abstract

Underwater images find application in various fields, like marine research, inspection of aquatic habitat, underwater surveillance, identification of minerals, and more. However, underwater shots are affected a lot during the acquisition process due to the absorption and scattering of light. As depth increases, longer wavelengths get absorbed by water; therefore, the images appear predominantly bluish-green, and red gets absorbed due to higher wavelength. These phenomenon’s result in significant degradation of images due to which images have low contrast, colour distortion, and low visibility. Hence, underwater images need enhancement to improve the quality of images to be used for various applications while preserving the valuable information contained in them.