Cloud Detection System for UAV Sense and Avoid: First Results of Cloud Segmentation in a Simulation Environment

The following paper provides cloud segmentation techniques as a functional component to enable unmanned aerial vehicles (UAV) to automatically detect and avoid clouds using electro-optical sensors. The motivation here is that UAVs not operating under air traffic control have to comply with visual flight rules (VFR). Therefore, minimum distances to clouds must be respected. This paper presents an approach to cloud segmentation in order to facilitate future investigations of cloud distance estimation. Apart from the system design, promising segmentation results are shown on synthetic images and subsequently discussed.