Underwater image restoration based on contrast enhancement

Abstract

Low contrast and color distortion are two major problems suffered by underwater images. In this work, a novel and systematic enhancement method based on contrast enhancement is proposed to restore the degraded underwater images. The proposed method mainly contains three steps. First, the quadtree subdivision strategy is adopted to estimate the waterlight. Then we formulate a cost function and minimize it so as to find the optimal transmission map which is able to maximize the image contrast. Once the waterlight and transmission map are obtained, based on the underwater image degradation model, the backscattering effect and color distortion caused by the light attenuation along the underwater object-camera path can be corrected. Finally, to further compensate for the color loss along the surface-object propagation path, a simple but effective color correction method is applied. Experimental results demonstrate that our proposed method can achieve comparable or even better results than some state-of-the-art approaches.