Calibration Measurements and Computational Models of Sensors Used in Autonomous Vehicles

An increasing number of vehicles are equipped with cameras. As perception sensors, they scan the surrounding area and supply the Advanced Driver Assistance Systems (ADAS) for building up an environmental model through the use of computer vision techniques. While they perform well under good weather conditions their efficiency is reduced by adverse environmental influences such as rain, fog and occlusion through dirt. As a consequence, the vision based ADAS obtains poor quality information, and the model also becomes faulty. This paper deals with methods to estimate information quality of cameras in order to warn the assistance system of possible wrong working conditions. In particular, situations of contamination or occlusion of the windshield or camera lens, as well as foggy weather are taken into account in this paper. In the issue of occlusion total, fractional and transparent effectuations have to be recognized and distinguished. Therefore, this paper proposes an approach based on edge analysis of consecutive frames and presents initial experimental results of the implementation. In the field of Fog Detection a method based on the Logarithmic Image Processing Model is described and the results are shown.