The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
approach using a combination of micro UAVs, consumer grade 
video cameras and virtual globes offers a point localisation or 
digitising accuracy in real-time between 6 and 15 meters. It 
should be further noted that these results are independent of any 
control points and that these accuracy figures apply on a global 
scale. 
a priori Geo-registration Accuracy 
Figure 7: A priori estimate for the achievable direct geo 
registration accuracy with the micro UAV 
'microdrones md4-200' 
The direct geo-registration accuracy achievable with low cost 
UAV platforms, video sensors and virtual globe technologies is 
more than adequate for many real-world applications. 
However, for certain application scenarios the above-mentioned 
geo-referencing accuracy of the direct geo-referencing approach 
is not yet sufficient. In such situations it is conceivable to use 
the available geospatial contents of the virtual globe to improve 
the geo-registration accuracy. This approach further denoted as 
integrated geo-referencing tries to continuously estimate and 
correct the systematic error part of the direct geo-referencing 
solution with image-based resection updates resulting from an 
image-to-model matching. First investigations of this geo- 
referencing approach show an accuracy improvement by a 
factor of four. A detailed accuracy investigation of the 
introduced direct geo-referencing and first investigations of the 
mentioned integrated geo-referencing approach performed with 
a similarly INS/GPS sensor configuration are presented in 
(Eugster and Nebiker, 2007). 
6. CONCLUSIONS AND OUTLOOK 
This paper presented a prototype solution for the integration of 
video imagery captured with mini or micro UAV systems into 
virtual globes - both in real-time and offline. Two different 
integration scenarios - the augmented monitoring scenario and 
the virtual monitoring scenario - were introduced. The 
augmented monitoring approach, for example, allows for the 
real-time 3D localisation or mapping of arbitrary geo-objects 
based on the video imagery content. A key element in the 
realisation of the video imagery integration is the video stream 
geo-registration process. This current video integration process 
is based on the direct geo-referencing approach, in which the 
achievable geo-referencing accuracy depends on the quality of 
the available flight attitude data and the synchronisation 
accuracy between video and flight data stream. A priori 
estimates and first field tests show that the presented solution 
offers a point localisation accuracy between 6 and 15 meters - 
anywhere on the globe and in real-time. This accuracy level is 
more than sufficient for numerous applications - especially 
when put into relation with the low costs and simple operation 
of our solution and with the fact that the system not only yields 
position information but also valuable context information in 
the form of real-time video views of the observed area and of a 
collaborative virtual environment. 
However, quite naturally the quality of the available flight 
attitude data of mini or micro UAV systems is much lower than 
that of the position and attitude data provided by the high-end 
GPS/INS systems used for direct geo-referencing in 
conventional airborne photogrammetry applications. In order to 
overcome these problems and to further increase the achievable 
mapping accuracy of mini or micro UAV systems, a new 
integrated geo-referencing approach has to be implemented. For 
such a solution an adaptive and robust image-to-model 
algorithm will be crucial. First investigations have shown that 
real-time image-to-model matching is a promising but very 
challenging task. Further, the matching process requires a 
robust video feature extraction which is delicate in outdoor 
environments due to changing weather conditions and variable 
lighting. 
REFERENCES 
Annen, A. and Nebiker, S. 2007: Einsatz von Mikro- und 
Minidrohnen fur Femerkundungsaufgaben in der 
agrochemischen Forschung und Entwicklung, Dreiländertagung 
der SGPBF, DGPF und OVG, DGPF Tagungsband Nr. 16, 
FHNW, Muttenz, pp. 637-647. (in German) 
Bento, M. 2008. Unmanned Aerial Vehicles: An Overview, 
InsideGNSS, January/February 2008, pp. 54-61. 
Bleisch, S. and Nebiker, S. 2006. Google Earth, NASA World 
Wind und Co. - wenn Geoinformation zum Modethema wird. 
Geomatik Schweiz, Februar/2006, (in German) 
Cramer, M., Stallmann, D. and Haala, N. 2000. Direct 
georeferencing using GPS/Inertial exterior orientations for 
photogrammetric applications; International Archives of 
Photogrammetry and Remote Sensing, v. 33, part B3, p. 198- 
205. 
Cramer, M. 2001. Performance of GPS/Inertial Solutions in 
Photogrammetry, in: D. Fritsch & R. Spiller, eds, 
'Photogrammetric Week '01', Herbert Wichmann Verlag, 
Heidelberg, pp. 49-62. 2001. 
Eisenbeiss, H. 2006. Applications of photogrammetric 
processing using an autonomous model helicopter, in: 
International Archives of Photogrammetry, Remote Sensing and 
Spatial Information Sciences, Vol. XXXVI-Partl/B, ISPRS 
Commission I Symposium, Paris, France, 03.-06. July 
Eugster, H. and Nebiker, S. 2007. Geo-registration of video 
sequences captured from mini UAVs - approaches and 
accuracy assessment, Mobile Mapping Technologies 
Symposium MMT 2007, Padua. 
1234