Underwater Image Restoration Based on Light Absorption

Due to the scattered lighting and light absorbed conditions, the underwater images frequently suffered from the low contrast as well as color distortion. As underwater images with the different color tones are captured under different light conditions, it is difficult to perform restoration and enhancement of such images. There are number of conventional methods reported to improve the quality of underwater images, however they constantly performs poor. In this paper, we propose a novel underwater image restoration and enhancement algorithm. We proposed the novel modified depth estimation approach for underwater images based on image blurriness as well as light absorption. To improve the processing speed as well as image quality we proposed novel image blurriness estimation approach in first we performed the image normalization and double precision conversion to improve the processing speed, then apply the edge-stopping pyramid technique rather than Gaussian filtering to improve image quality. The proposed approach is simulated and evaluated against the state-of-art methods in terms of processing time, PSNR, SSIM etc.