The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
1052 
In the third part, the results of the test investigation of the 
system accuracy are presented. Here main focus is laid on the 
analysis of the Dual-DigiCAM-H/39 geometric accuracy 
potential, obtained from the analysis of independent check point 
differences. 
The accuracy of the direct georeferencing using the 
AEROcontrol GPS/IMU results are also analysed. 
2. SYSTEM SETUP 
The operated Dual-DigiCAM-H/39 consisted of two 39 Mpixel 
medium format digital cameras (Grimm & Kremer 2005). They 
were rigidly connected to ensure a fixed relative orientation in- 
between the two cameras and compared to the attached IMU. 
The camera on the left side in flight direction was mounted with 
a roll angle of -14.8° (pointing to the right side) and the camera 
on the right was mounted with a roll angle of +14.8°. For the 
used 82.3 mm lenses and the format of 7216*5412 pixels for the 
single cameras, these mounting angles result in an image 
overlap of about 782 pixel or 10.8%. With the given image 
width of 7216 pixel, this configuration gives an effective image 
width of 13650 pixel. Both cameras were mounted with the 
short side of the image (5412 pixel) in flight direction. 
(“GSD”). The planned average ground sample distance for the 
different blocks was planned to be 7cm, 14cm and 20cm, 
respectively. 
Figure 2: System installation in the aircraft. 
Figure 1: Arrangement of the two cameras and the IMU in the 
sensor pod. 
'V 
The camera was operated together with an AEROcontrol-IId 
(256Hz) GPS/IMU system for the precise determination of 
position and orientation of the camera at the instant of exposure. 
The camera system, consisting of the Dual-DigiCAM and the 
AEROcontrol IMU was mounted in a GSM-3000/IGI stabilised 
sensor mount from SOMAG AG, Jena, Germany. The real time 
orientation information of the AEROcontrol was fed into the 
stabilised mount to improve the levelling accuracy. 
For flight guidance and sensor management, a CCNS4 
navigation system was operated. 
a Coctroi Points *■ Cfack Point* 
Figure 3: Top: Block layout GSD 7cm. Middle: Block layout 
GSD 14cm. Bottom: Block layout GSD 20cm. 
The 7cm block included three east/west lines with 27 double 
images each and three cross strips with 13 double images. The 
overlap was p=60% in forward direction and q=76% in cross 
flight direction. The 14cm block had three lines east/west with 
3. FLIGHT PLANNING AND DATA COLLECTION 
The flight mission consisted of three sections. The sections 
were image blocks with three different ground sample distances