X-axis parallel to the direction of flight. The object points
are arranged in a grid with AX —12.5 km, AY = 12.5 km
and Z —0km. Consequently the single strip consists of
260 and the blocks contain 1,092 object points. For each
block configuration, the image coordinates of the object
points were computed assuming a straight forward flight
path with a constant altitude of 400 km and attitude val-
ues equal to zero. All image coordinates were treated as
being uncorrelated with equal standard deviations of 0.3
pixel.
d) Ground Control Information
Either no or 16 GCP with ox = oy = oz = 1.0 m were
used. The 16 GCP are arranged in 4 groups of 4 points
each, located at the corners of the 3-ray area of the strip
or the block. The standard deviations of the image coor-
dinates are assumed to 0.5 pixel.
e) Orbit and Attitude Observations
For the orbit and attitude observations 3 different cases
were investigated (Tables 3 and 4). Case B is the most
realistic one. In Table 3 the standard deviations describing
the relative accuracy of the position parameters are zero,
since all camera positions are constrained to lie on the
orbit trajectory. The standard deviations describing the
absolute accuracy are 5 m for the position and 2cm/s for
the velocity components of the epoch state vector.
Position
A B C
relative 0m 0m 0m
position 0m 5m 20m
beate velocity | 0cm/s | 2cm/s 8cm/s
Table 3: Standard deviations of the observed position pa-
rameters
Attitude
A B C
relative 0” 10" 2
bias 0^ 200" no obs.
absolute drift | 0”/s 0.7"/s no obs.
Table 4: Standard deviations of the observed attitude pa-
rameters
For each orientation point attitude observations were in-
troduced with a relative accuracy of 10” and an absolute
accuracy of 200" (bias) and 0.7" /s (drift) respectively (Ta-
ble 4). Besides case B, two extreme cases are investigated
for comparison: error-free (A) and worst case (C) orbit and
attitude observations. Based on experiences with MOMS-
02/D2 data, the distance between the orientation points
was chosen to 4,940 rows (ca. 90 km) leading to 9 orienta-
tion points per strip.
5.2 Results
For analysis, the rms values p4¢ (planimetry) and pj
(height) of the theoretical standard deviations 04, oy and
162
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996
c; of all 3-ray points (single strip), 3- and 6-ray points
(block with q-— 2096) as well as 3-, 6- and 9-ray points
(block with g — 6096) were calculated. Moreover, the rms
values po, fp, pi of the theoretical standard deviations of
the estimated exterior orientation parameters &, Q, & at
the orientation points were computed. All accuracy figures
were derived from the inverted normal equation matrix.
Figure 4: Rms values pg, for different block configura-
tions, ground control information and orbit/attitude data
RMS VALUES B2) [m]
Figure 5: Rms values pn; for different block configurations,
ground control information and orbit/attitude data
In Figures 4-8 these rms values are shown graphically,
where A,B and C stand for the standard deviations of the
observed position and attitude parameters according to
Tables 3 and 4.
First the results of the single strip adjustments without
GCP are discussed. Assuming error-free position and at-
titude data
depends or
nates, the
constellatic
3 m in plan
Figure 6: |
ground coi
Figure 7: -
ground co
'The plani
sition and
dard devi:
z 077 m e
true to Sc
unfavoura
(bias) anc
eters w. I
