Abstract—Nowadays cars are equipped with more and more sensors to be able to drive autonomously. Soon cars will become service robots which require a precise self localization. But instead of interacting in a goodwill environment like offices, hallways or warehouses, automated cars will be interacting in the real world. Hence they have to handle different types of challenges like seasonal and structural change, changing lighting, and weather conditions. These atmospheric conditions influence the required precise self localization. Therefore this paper introduces a method to generate weather specific field of view parameters for dashed road marking landmarks extracted by a mono camera. With these parameter sets a useful reduction of the measurement set is achieved. This should lead to an improvement of the localization accuracy compared with standard field of view parameters. To this end, we examine three atmospheric conditions: rain, wetness and dryness.

Index Terms—localization, fuzzy learning, landmark based localization, field of view, atmospheric conditions, weather conditions

Cite: Marek Steß, Michael Schlichte, and Bernardo Wagner, "Camera-Based Field of View Parameter Optimization," International Journal of Mechanical Engineering and Robotics Research, Vol.4, No. 4, pp. 293-298, October 2015. DOI: 10.18178/ijmerr.4.4.293-298