A Gray Scale Correction Method for Side-Scan Sonar Images Based on GAN

During the acquisition of a side-scan sonar image, energy loss occurs in the sonar acoustic waves, which lead to gray scale distortion in side-scan sonar images. Therefore, gray scale correction of the side-scan sonar image is necessary before further processing of side-scan sonar images. In this paper, we introduce the causes of gray scale distortion in side-scan sonar images and several commonly used methods of image gray scale correction. Analyze the shortcomings of existing methods, according to the characteristics of the side-scan sonar image, we propose a gray scale correction method based on generative adversarial network(GAN). In addition to presenting a new approach, we introduce a new refined loss function for achieving improved results. The loss function is aimed at reducing artifacts introduced by GAN and achieve better visual quality. Experiments show that the proposed method can effectively correct the gray distortion in the image.