KNMI – Fog Detection from Camera Images #SWI2016
KNMI products and services rely heavily on observations. Besides increasing amounts of data from traditional sources (such as observation networks), the use of open data, crowd-sourced data and the Internet of Things (IoT) is rapidly emerging. To deploy these sources of data optimally, the KNMI DataLab has been established. Data‐driven innovations that arise from public‐private collaborative projects and research programs can be explored and facilitated by the KNMI DataLab.
Fog is a hazardous type of weather. The impact of fog has significantly increased as a consequence of the increase in air, marine, and road traffic. The loss of life as a result of fog can be compared to the losses of other weather events, such as tornadoes or even hurricanes. Fog can form suddenly and can dissipate just as rapidly. Visibility/Fog can be measured using so‐called Transmissometers or Forward‐scatter‐visiometers. These sensors are mainly used along the runways of airports and are sparsely distributed over the world.
An alternative might be to use camera images, as it is quite simple for the human eye to see the differences in the two pictures on this page.
Combined with the vast amount of image data, e.g. from webcams or traffic cam data, automatic fog detection from cameras is potentially very interesting for the KNMI DataLab. Therefore, we would like to ask you, which kind of algorithms are best fitted for automatic fog detection?
We are particularly interested in algorithms, that can be adapted easily to different cameras. For the SWI we provide camera data from the KNMI measurement field, together with other relevant meteorological data.