Underwater image enhancement algorithm combining color correction and multi-scale fusion

Abstract

Due to the absorption of light by water and the scattering effect of suspended particles on light, underwater optical images often have many problems, such as noise interference, blurred details, uneven illumination, low contrast and color distortion. In order to solve those problems, a novel underwater image enhancement algorithm which combines color correction and multi-scale fusion is proposed in this paper. First, the color of the underwater images are corrected by the improved multi-scale retinex with color restoration algorithm. After that, the contrast and edge details of the color-corrected image are enhanced by the contrast limited adaptive histogram equalization algorithm and sharpening algorithm. Then, we calculate multiple weights of the contrast-enhanced image and edge-enhanced image to construct normalized weight maps. Finally, we input the images into the Laplacian pyramid, and input the normalized weight maps into the Gaussian pyramid for multi-scale decomposition and fusion reconstruction layer-by-layer . Through multi-scale fusion, we combine the advantages of color-corrected image, contrast-enhanced image and sharpened image without introducing Manmade noise. Comparative experiment results show that the underwater image enhancement algorithm proposed in this paper can significantly improve the contrast of the underwater images, correct the color deviation, enriches the edge details and make the image clearer.