Underwater haze removal using contrast boosted grayscale image

Abstract

Underwater images generally experience the ill effects of color shift, haze and contrast deterioration because of scattering and absorption of light in water. There are several techniques that can increase the contrast of an image but most of the techniques lose the originality of the image while increasing the contrast, making these techniques highly inefficient and undesirable. The main goal of the proposed work is to increase the contrast of an image along with the haze removal while keeping the original color intact. A two-step decolorization process is employed for the contrast boosting of the grayscale image, which is later used for reproducing the enhanced color image. Mean and standard deviation-based contrast mapping is used to increase the global contrast and Laplacian pyramids are used for enhancing the local contrast adaptively by identifying the edges in an image. Finally, by combining both the local and global approach a contrast boosted image is produced and the color image is retrieved by changing the L component in CIE Lab color space.