Image dehazing method by fusing weighted near-infrared image
Yuhei Kudo, Akira Kubota · 2018 International Workshop on Advanced Image Technology (IWAIT) · 2018
We propose a novel color image dehazing method based on fusing a visible and a near-infrared (NIR) images of the same scene. In outdoor scenes, image contrast of the captured images is likely to be lost by haze. Near-infrared light is less scattered than visible light because of its long wavelength; therefore the NIR images have the contrast enough. Our approach restores the image contrast by fusing detail components of the NIR image into the visible image. To prevent overemphasizing haze-free regions by NIR information, we introduce weighting the NIR image using the transmission map. Experimental results show that our method outperforms conventional methods.