Underwater image enhancement by the combination of dehazing and color correction

Abstract

Underwater image processing is crucial for many practical applications in the ocean filed, which is not a trivial thing since the environment of underwater is often complicated and short of light. The major difficulty is that a captured image is fuzzy, under-exposed and often has the color cast due to the fact that the light is absorbed and scattered. To overcome those difficulties, we propose a new underwater image enhancement method, which is composed of two successive vital processings: the dehazing and color correction. Firstly, considering the characteristic of light propagation under water, we propose a new dehazing algorithm to restore the visibility of degraded underwater images based on the dark channel prior, through building up the relationship of the transmission rates among three color channels. Then, to further improve the image quality, we adopt an effective color correction method on the obtained haze-free underwater images. We conduct extensive experiments under measure tests of both subjective and objective, and the results show that our method is superior to several existing approaches.