A passive cloud detection system for UAV: Weather situation mapping with imaging sensors

This paper presents a passive cloud detection system on-board of unmanned aerial vehicles (UAVs). Clouds are to be avoided especially by lightweight UAVs to ensure structural health, as their presence indicate weather phenomena like turbulences and icing. For larger airplanes, local weather observation is traditionally achieved by weather radar. Since such active sensors with their high energy consumption and mass cannot be used on-board lightweight UAVs, the proposed cloud detection system is based on a passive imaging sensor in the visual spectrum (EO) and computer vision methods. For this work a High Altitude Pseudo-Satellite (HAPS) serves as example platform. Due to its relatively low airspeed and maneuverability, reactive maneuvers are best to be avoided. For ascent and descent through cloud layers therefore, a sufficiently large cloud-free corridor has to be identified beforehand. This is achieved by utilizing two modes of operation, which are described in this paper. The steps are demonstrated in a virtual environment example with focus on mapping and integration of the gathered information into a general weather situation map.