An effective and robust underwater image enhancement method based on color correction and artificial multi-exposure fusion

Underwater images/frames are always subjected to color distortion, contrast reduction and detail loss, which degrade the visual quality severely. Current dehazing methods could not improve the visual quality of underwater images/frames robustly and effectively, especially in removing the undesired color cast. To address the issue, this paper introduces an effective and robust underwater image enhancement method without any dedicated hardware or prior knowledge. First, an adaptive reduction operation on the two stronger color-channels of inputs is employed to avoid the red over-compensated deficiency appearing in color-balanced result. Second, three kinds of color-balanced images are generated from the operation, which combines color compensation algorithms and famous Gray-World assumption. Third, a novel algorithm based on two non-reference quantitative evaluation indicators is utilized to choose the optimal color-balancing version. Then, gamma adjustment operation is employed to generate artificial over-/under-exposure visions of color-balancing image. Last, 'exposedness' and 'contrast' are set as two weights, being blended into the famous multi-scale fusion framework to generate the enhanced result. Our experimental results demonstrate the superior performance of the proposed method in both subjective and objective evaluations. Besides, the proposed method is also suitable for dehazing regular fogged images and local feature points matching.