Single image dehazing using the change of detail prior

In this paper, we present a simple but effective image prior, change of detail (CoD) prior, to remove haze from a single input image. The CoD prior is developed based on the multiple scattering phenomenon in the propagation of light. By adopting the CoD prior in an image model, the thickness of the haze can be estimated effectively to recover a high quality, haze-free image. Our method is stable to image local regions containing objects in different depths. In addition, since the CoD prior is mostly sensitive to local details of the image rather than color and intensity, our method is able to handle both color and grayscale images. Our experiments showed that the proposed method achieved better results than several state-of-the-art methods, and it can be implemented very quickly.