Underwater image restoration using color correction and non-local prior

Underwater images often suffer from color and contrast degradation, because the light is absorbed and scattered while traveling in water. Although the physical process of the underwater images seems similar to the outdoor haze images, conventional dehazing methods fail to generate accurate results since colors associated to different wavelengths have different attenuation rates in underwater conditions. To overcome this, we propose a novel underwater image restoration method based on color correction and image dehazing. First, we estimate the global background light using a hierarchical search based on quad-tree subdivision combined with the ocean optical properties. According to the properties of underwater optical imaging, we then introduce an underwater color correction method using depth compensation, in which a multi-channel guided image filter is proposed to refine the depth image. Finally, we adopt the non-local image dehazing algorithm to restore the underwater images. Experimental results demonstrate that the restored images can achieve better visual quality of underwater images when compared with several state-of-the-art methods.