Improving the Contrast Enhancement of Oceanic Images Using Modified Dark Channel Prior

Oceanic photography faced a difficult situation that the light under the water was not only scattered but also absorbed by the medium at the same time. The typical behaviour in an underwater image contains two parts: First, it usually had three regions with different colour (usually yellow, green and blue). It depended on the depth or the distance from the supplement light source. Second, the contrast was attenuated due to the scattering caused when the light passed through water molecule. Although there were researchers revealed that the haze removal algorithms such as the Dark Channel Prior improved the contrast and saturation loss during the capturing, there were only some problems solved so that further colour correction or compensation was indispensable. The primary disadvantage that adopted the haze removal method was that this algorithm usually based on the assumption that the atmospheric light was achromatic or uniform distribution so that the colour bias would not truly corrected. This research proposed a modified version of the Dark Channel Prior to better fit the condition under the water.