Underwater Image Restoration Using White Balance and Retinex Algorithm

Abstract

When light travels through water, attenuation and scattering of light have a negative impact on underwater vision. Additionally, this negative impact affects the underwater image quality and creates a fuzz problem. An innovative retinex-based enhancing method is suggested in this study to improve a single underwater image. The proposed method consists of three phases: white balancing approach, decomposition of an image into illumination and reflectance using the retinex-algorithm, and enhancing the illumination and reflectance. The enhanced illumination and reflectance multiply to obtain an improved y-component of YCbCr color space. Finally, concatenate all components of YCbCr color space and convert it into RGB color space to get an enhanced image. The author evaluates metric parameters to validate the proposed method.