Underwater Image Enhancement by Fusion

Abstract

Underwater images are suffered by several factors like color absorption, light scattering, and these lead to poor visibility and low contrast of an underwater image. Typical image restoration methods rely on the popular Dark Channel Prior. However, the result of the underwater image which has only been dealt with the restoration algorithm is still not ideal. In order to get better results, we introduce a new underwater image enhancement approach based on multi-scale fusion strategy in this paper. In our method, we first obtain the restored image on the base of underwater image model. Then we get the white balance and contrast enhancement image of the restored image respectively. Finally, these two derived inputs are blended by multi-scale fusion approach, using saturation and contrast metrics to weight each input. This algorithm reduces the execution time and can effectively enhance the underwater image. The experimental results demonstrate that our method can obtain better visual quality.