Fog degree measurement based on local contrast and color similarity

In the single image-based fog related systems, prior information on the degree of fog is essential. In order to quantify the degree of fog in a single image, we propose a method that not only measures the sizes of regions in which the light transmission values, are expected to be foggy ones, but that also determines the range of local contrasts in the expected foggy regions. The transmission is estimated from the dark channel prior method. To prevent the proposed method from increasing in complexity compared to the dark channel prior method, we devise a method which calculates the local contrast during the dark channel prior procedure. The proposed method quantified the degree of fog with more than 95 percent accuracy for images containing various objects and environments. Furthermore, processing time of the proposed method is fast enough to be applied to be real-time system.