Single Image Fog Removal Based on Dark Channel Prior and Local Extrema

In this paper, we propose a novel atmospheric model–based defogging method from single image, for the image degraded by scattering due to atmospheric particles. The main advantage of the proposed algorithm compared with previous techniques is that it can reduce the Halo and blocking effects physically. We estimate the transmission using local extrema principle and bilateral filter that a high quality depth map can be obtained. Experiments on a variety of outdoor haze images demonstrate that our algorithm can restore well colors and contrast of the observed objects, improve the visibility of image and avoid the Halo and blocking effects effectively.