Mathematical description of the fine structure of edge areas of vehicles for traffic control

For the detection of traffic flow parameters and the traffic state, different types of elementary properties within traffic images, especially around the edges, are selected and fused. Textural properties contained in edges of vehicles or their surroundings are robust against change of lighting. These edge areas are only a small part of the. image and can analysed in real time. Exploring the structure of edge areas gives the possibility to divide partially occluded vehicles and to estimate the height by analysing and comparing parts of vehicle edges. The properties across the edges are rapidly changing and need new mathematical description. The isolated and interrupted properties of textural, structural, radiometrical and geometrical kind are described by the Lebesgue measure. These measures are fused by the Lebesgue integral. If the knowledge is incomplete, fuzzy measure and fuzzy integrals are used. By fusion, characteristic properties for the vehicles are generated and used for the detection of vehicles.