Figure 12. Snapshot of Geofort DSM 
1262 
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
The rotorcrafit UAV of Geocopter was asked to produce the 
actual DSM and orthophoto of the Geofort. The system flew in 
fully automatic mode over this site in April 2008 and captured 
the airborne photogrammetric imagery in four parallel flight 
lines at 100 meters above ground level. Table 10 lists the 
project parameters of the UAV survey flight. 
3.3 Corn field, Switzerland 
This project was conducted to determine to what extent 
topography can affect the rate of cross-pollination. The effect 
of topography on cross-pollination was investigated in 2005 and 
2006 in canton Zurich. 
Parameter 
Value 
Flying height 
100 m 
Flight lines 
4 
Photos 
80 
Ground Control Points 
20 
Forward lap 
60% 
Sidelap 
30% 
Project area size 
500 x 300 m2 
Footprint one photo 
80 x 60 m2 
Pixel size 
2 cm 
Camera 
Fuji FinePix Pro S5 
Lens 
Nikon 28 mm 
Time of survey 
12 minutes 
Weather 
high clouds 
Wind 
fair 
Table 10. Geofort Project Parameters 
Figures 11 and 12 display the level of detail of the aerial 
photography and the extracted DSM points from Match-T. 
UAV survey flights were conducted over two sites (longitude 8° 
40’ E, latitude 47° 25’ N, flying height about 500 and 450 
meters above mean sea level) (Eisenbeiss, 2007 and 2008). In 
2005, all imagery over the com field was acquired in 20 
minutes under autonomous flight with stop points. The time of 
flight was reduced to 5 minutes in 2006 under autonomous 
cruise mode. The obtained images were oriented with software 
packages LPS (Leica Photogrammetry Suite 9.0, Leica 
Geosystems Geospatial Imaging, LLC) and ISDM (Image 
Station Digital Mensuration, Zeiss Intergraph, 2000). The 
aerotriangulation of the UAV images was performed with a a 0 
of less than 0.5 pixels and an RMSE for the control points of 
less than 1 dm for both test areas. 
A 10-cm GSD elevation model was generated with SAT-PP 
software (Satellite Imagery Precision Processing, ETH Zurich, 
Zhang, 2005), that was modified to also process imagery from 
calibrated small format digital cameras. At last, the orthophoto 
with 3-cm resolution was produced with LPS. The accuracy of 
the results was mainly affected by difficulties in the 
aerotriangulation, which was because of the high repetitive 
texture in the com fields. Figure 13 illustrates the matching 
result with SAT-PP: feature points and edge matching. 
The generated DSMs and orthophotos were together integrated 
in a GIS, and used for high resolution measurements on the 
outcrossing in com. Figure 14 shows the orthophoto draped 
over the DSM (both from 2005), combined with the Swissimage 
orthoimage (swisstopo®) draped over the DHM25 
(swisstopo®), and the position of the cross-pollination data. 
The accuracy of the DSM was not possible to verify since 
reference data of the com field was not available. However, the 
results from the orientation of both years are quite consistent; 
therefore the DSMs of both years have the same accuracy level.