Underwater image restoration based on depth map

The estimated accuracy of the depth maps of underwater images affects the quality of the restored images. In order to obtain more precise depth maps, an algorithm was proposed to calculate the depth map based on attenuated channels and luminance map, and then the depth map was used to recover underwater images. First, the depth map of underwater image is estimated according to the relationship between image pixel and scene depth. Then the depth map is further rectified and refined by using the luminance map. Third, the atmospheric light value and the transmission map of the image is calculated using the refined depth map. Finally, the degraded underwater image is restored by inversely solving the underwater imaging model. The experimental results show that compared with the existing algorithms, the model parameters calculated using the rectified depth map are more accurate, and the restored image has better contrast and can maintain more natural color. © 2021, Jilin University Press. All right reserved.