A vision based landing spot searching for unmanned aerial vehicle using satellite image

Recently, unmanned smart devices have become popular in various fields from the military to civil use. Flight in complex urban areas requires high-level autonomous flying techniques to ensure the safety of people. In addition, UAVs can encounter critical circumstances when flying because of inevitable obstacles, component malfunction, or faulty systems. These factors can force emergency landing on suitable spots. This paper proposes an emergency landing procedure that uses pre-selected available landing spots from satellite images, and navigates UAVs to such spots in an emergency. Searching suitable landing spots using a camera in an emergency is time-consuming. This paper also proposes a simple method that searches for landing spots in satellite images using proper shapes and areas. By applying this method, our results estimate that appropriate landing spots can be searched correctly.