Bachelor thesis – 3D-Objektverfolgung mit Stereokameras zur bildbasierten Navigation autonom fliegender Luftfahrzeuge

Nowadays, unmanned aerial vehicles (UAV) are used in many scenarios. Since a collision or even a crash would be fatal, it is important for a UAV to be able to navigate even in an unknown environment. To achieve this, camera systems, especially stereo camera systems can be used to optain an optical navigation solution. This bachelor thesis’ goal is to develop an algorithm to detect and track homologous points in a stereo image series over time. Using stereo-geometric calculations it is possible to calculate the position of those features in three-dimensional space. These points and vectors can be used to calculate relative position changes of the vehicle itself or in further projects to create an obstacle map of the environment. The results are validated using image data from a manual flight test on measured reference objects.

Paper – Stereoskopische Bewegungsschätzung zur optisch gestützten Navigation von unbemannten Luftfahrzeugen

F. Andert, F. Bäthge, J. Dittrich
61. Deutscher Luft- und Raumfahrtkongress, Berlin, Deutschland