Underwater Image Restoration Method Combining Improved Red Channel Prior and Power Law Correction-based CLAHE Algorithm

An underwater image restoration method was proposed to address the issues of background light estimation deviation and contrast imbalance that occurs during the restoration of water degradation images. First, the background light area and value are determined according to the superpixel image segmentation method, and then the red channel prior theory is used to obtain the estimated transmittance and the preliminary restored image. Finally, the color of the restored image is enhanced using the normalized power law correction-based contrast limited adaptive histogram equalization (CLAHE) algorithm. Three image quality evaluation standards are used to objectively analyze the experimental results, and it is found that the proposed method can effectively balance the contrast and improve the visualization effect.