Fusion Strategy for Single Image Dehazing
Fan Guo, Jin Tang, Zixing Cai · International Journal of Digital Content Technology and its Applications · 2013
Recovering the image captured in bad weather remains a challenging task due to the inherent ambiguity. In this paper, we propose a fusion strategy that derives the inputs and weight maps only from the original degraded image. The two inputs represent color corrected and contrast enhanced versions of the original hazy image, and the three weight maps aims to increase the visibility of the scene. Our strategy is a single image method that does not require physical model or scene structure. Besides, a novel indicator for dehazing effect assessment is also proposed from the perspective of human visual perception. We experimentally evaluate our algorithm on a wide variety of hazy images, the qualitative and quantitative results demonstrate that the proposed method is effective and robust, yielding high-contrast and vivid haze-free images.