ertical 
ssical 
ntage 
of tie 
or a 
juired 
‚if no 
th the 
bined 
s. Ina 
by a 
nd so 
ferent 
case, 
ey do 
netry 
lateral 
(d use 
'S. AS 
i. The 
mined 
| - the 
. The 
icient 
^ction 
ina is 
| to be 
ach a 
ay the 
1 area 
h 252 
U, so 
ct 
CR88 
2000. 
rcraft, 
f the 
ction 
of 
arrain 
th of 
. The 
angle 
| the 
mera, 
cy is 
d the 
nt is 
Karsten Jacobsen 
  
  
influenced by the more strong correlation of the orientation unknowns. The large photo scale was required for the 
identification of the mapping objects and not for the accuracy. An accuracy of +/-1m in X and Y was required for the 
ground points corresponding to 285um. But for a stereoscopic view, the y-parallax should not exceed 30pm. 
Another small block has been processed in cooperation with the University of Applied Sciences, Bochum, Germany 
(Bäumker et al 1999). With a self developed, fast reacting stabilized platform for carrying a digital camera Kodak 520, 
the sensor orientation has been determined and so the camera was very close to a nadir view. 
3 PREPARATION OF THE INERTIAL DATA 
The IMU is determining roll, pitch and yaw. Also after correction by the boresight misalignment such rotations cannot 
be used directly for the setup in photogrammetric work stations. The primary rotation yaw is related to geographic 
North instead of the grid North used in photogrammetry. With the program IMUPRE of the Hannover program system 
BLUH the IMU-attitude data have been converted into the usual photo orientations respecting the convergence of 
meridians, the different rotation definition and the dimension of the attitude data (grads instead of degrees). By a 
comparison of the photo orientations of the reference blocks with the orientations determined by means of GPS and 
IMU, the relation of the axis between the photogrammetric camera and the IMU has been determined as well as 
systematic differences of the GPS-positions. By linear time depending interpolation, based on the results of the 
reference strips flown before and after the flight over the main area, the photo orientations of the images in the main 
area have been improved. The improvement of the attitude data was done in the pitch, roll and yaw-system, 
corresponding to the relation of the axes. 
The photo orientation determined by bundle block adjustment is not free of errors, especially in the case of a small view 
angle the orientation elements do have strong correlation’s. Especially the correlation of X0 and YO to phi and omega is 
in the range of 0.995, that means, shifts of the GPS-data and angular misalignment cannot be separated totally (Jacobsen 
1999). The separation of the components can be improved by a flight over the reference area in opposite direction, but 
this was not done. 
In general a problem of the separation will only have a negative influence if the flight over the main block area will be 
made under different conditions. In the case of the same image scale and same flight direction, a separation of the 
components is not required. For the fast reacting platform, equipped with a Kodak DCS 520, the view angle was only 
24° x 34.7°, causing a correlation of the orientation elements computed by resection up to 0.999. It was not possible to 
separate shifts in X0 and YO from roll and pitch. By this reason, after an approximate shift of the GPS-data, the 
projection center coordinates X0 and YO have been fixed for the determination of phi and omega. 
bo t 
roll itch 
tic differences da -.445 -.469 
« da -.454 -.462 
day 3 -.463 -.462 
da -.477 -.471 
differences 
without matic differences da .039 .012 .044 
ke da ‚029 .016 .049 
day : .042 .021 17 
“ da ‚034 015 .091 
er linear fitti 025 .009 .007 
« .021 .009 .010 
.026 .014 .012 
.017 .016 .008 
da 011 .009 .007 
da .021 .009 .010 
day 3 018 015 011 
da 018 006 .005 
Table 1 : differences of the attitude data IMU - controlled bundle block adjustment (reference flight strips) 
In table 1 the attitude differences between the controlled bundle block adjustment and the IMU-data in the reference 
area are shown. The misalignment in roll and pitch is constant over the 4 days within the standard deviation. This is not 
the case for yaw. A significant change of the yaw between the flight over the reference flight strip before the main area 
to after this happened (much smaller values "after linear fitting"). If this change is a linear function of the time, it can be 
respected in the determination of the image orientation based on the IMU-data. The determination of the misalignment 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000. 431