Volume XLII-2/W6
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-2/W6, 85-92, 2017
https://doi.org/10.5194/isprs-archives-XLII-2-W6-85-2017
© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 4.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-2/W6, 85-92, 2017
https://doi.org/10.5194/isprs-archives-XLII-2-W6-85-2017
© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 4.0 License.

  23 Aug 2017

23 Aug 2017

UAV CAMERAS: OVERVIEW AND GEOMETRIC CALIBRATION BENCHMARK

M. Cramer1, H.-J. Przybilla2, and A. Zurhorst3 M. Cramer et al.
  • 1Institute for Photogrammetry (ifp), Stuttgart University, Geschwister-Scholl-Str. 24D, 70174 Stuttgart, Germany
  • 2Department of Geodesy, Bochum University of Applied Sciences, Lennershofstr. 140, 44801 Bochum, Germany
  • 3aerometrics, Landwehrstraße 143, 59368 Werne, Germany

Keywords: UAV camera, geometric calibration, stability, benchmark, image formats

Abstract. Different UAV platforms and sensors are used in mapping already, many of them equipped with (sometimes) modified cameras as known from the consumer market. Even though these systems normally fulfil their requested mapping accuracy, the question arises, which system performs best? This asks for a benchmark, to check selected UAV based camera systems in well-defined, reproducible environments. Such benchmark is tried within this work here. Nine different cameras used on UAV platforms, representing typical camera classes, are considered. The focus is laid on the geometry here, which is tightly linked to the process of geometrical calibration of the system. In most applications the calibration is performed in-situ, i.e. calibration parameters are obtained as part of the project data itself. This is often motivated because consumer cameras do not keep constant geometry, thus, cannot be seen as metric cameras. Still, some of the commercial systems are quite stable over time, as it was proven from repeated (terrestrial) calibrations runs. Already (pre-)calibrated systems may offer advantages, especially when the block geometry of the project does not allow for a stable and sufficient in-situ calibration. Especially for such scenario close to metric UAV cameras may have advantages. Empirical airborne test flights in a calibration field have shown how block geometry influences the estimated calibration parameters and how consistent the parameters from lab calibration can be reproduced.