Fog detection from camera images

Fog is one of the most dangerous weather types with more fatalities than\u3cbr/\u3ewinter storms. It is in the interest of general public that a precise, predictive\u3cbr/\u3eand accurate fog density map with high spatial resolution can be created. Currently, the definition of fog as used by national weather services is so detailed and technical that the fog can be identified only at a few locations by means of the prescribed light scattering experiments. With the rising availability of cameras in public places such as airports, streets and highways, a large amount of data on the occurrence of fog becomes available to researchers. In this article we describe methods for determining not necessary only the existence of fog, but sometimes a visibility distance - a type of optical penetration length - as well. We will show that digital cameras can be a reliable alternative or complementary method for creating fog visibility maps when processing of image data is used