Investigations on the quality of the interior orientation and its impact in object space for UAV photogrammetry

Abstract

With respect to the usual processing chain in UAV photogrammetry the consideration of the camera's influencing factors on the accessible accuracy level is of high interest. In most applications consumer cameras are used due to their light weight. They usually allow only for automatic zoom or restricted options in manual modes. The stability and long-term validity of the interior orientation parameters are open to question. Additionally, common aerial flights do not provide adequate images for self-calibration. Nonetheless, processing software include self-calibration based on EXIF information as a standard setting. The subsequent impact of the interior orientation parameters on the reconstruction in object space cannot be neglected. With respect to the suggested key issues different investigations on the quality of interior orientation and its impact in object space are addressed. On the one hand the investigations concentrate on the improvement in accuracy by applying pre-calibrated interior orientation parameters. On the other hand, image configurations are investigated that allow for an adequate self-calibration in UAV photogrammetry. The analyses on the interior orientation focus on the estimation quality of the interior orientation parameters by using volumetric test scenarios as well as planar pattern as they are commonly used in computer vision. This is done by using a Olympus Pen E-PM2 camera and a Canon G1X as representative system cameras. For the analysis of image configurations a simulation based approach is applied. The analyses include investigations on varying principal distance and principal point to evaluate the system's stability.