Underwater optical image processing based on double threshold judgements and optimized red dark channel prior method

Underwater images are prone to suffer from color distortion and low visibility because of the strong light attenuation. The traditional dark channel prior (DCP) method tends to fail when used for underwater image restoration. By exploring the differences in light attenuation between atmosphere and water, we propose an innovative image restoration method-optimized red dark channel prior (ORDCP) which adds a valid contrast indicator. In addition, we set double threshold judgments to determine the main color tone and calculate red channel transmission map. After getting the two estimated parameters including transmission maps and background light, we can restore underwater images using conventional underwater imaging model. The subjective evaluations indicate that the algorithm we proposed has better performance in terms of saturation, contrast and images edge details. What's more, the results of objective evaluation metrics show that the performances maximally increase by 32.32% in underwater images such as the ship and stone. The conclusion can be drawn that the proposed method is able to remove the noise and blur caused by complicated underwater environment and performs favorably against the state-of-the-art algorithms.