An Efficient Color Correction based Fusion Method for Underwater Image Enhancement

Abstract

In the marine industry, vision computing techniques are used in various domains like mine detection, underwater microscopic detection, and exploration of marine life and its protection. The captured underwater images mainly suffer from color distortion, low contrast, poor illumination, and of blurred details due to the optical properties of water. To improve the visual information and accuracy of recognition, a fusion method for underwater enhancement is proposed. The two versions of the input images result from color correction and dynamic histogram equalization and are fused using a multiscale fusion process to get an enhanced image with the natural color tone, and enhanced its finer details of the image. The qualitative and quantitative analysis of the proposed method yields excellent results for several datasets.