International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004
E 40 T
A
S os
=
3
Sos
=
= |
'e 04 1
a i
x ........ .. GSD
S 0.2 .. se
5 i
o |
5 0.0 ; E =
[2] Japan Lakeland NASA Lakeland Ajax 03
= Feb03 Dec02 Stennis Mar04
BRMS X BRMS, Y CIRMS Z
Figure 8: Statistics of Checkpoint Residuals for Different
DSS Test Flights — 55 mm Lens
Each of the test flight data sets discussed here is processed
similarly by following the data flow represented in Figure 2,
where the DSS image data is radiometrically balanced while the
POS AV GPS-aided inertial data is processed. Then the image
data, the exterior orientation. parameters, and the system
calibration parameters are run through an airborne calibration
and quality control procedure. Then the data is used to produce
ground coordinates of the available checkpoints.
The resulting coordinates are then compared to those computed
independently using land surveying techniques. The resulting
checkpoint residuals are listed in Table 2, while the RMS
values are shown in Figure 8 and Figure 9.
Flight Statistic parallax ih av az
(um) (m) (m) (m)
Min 0.2 0.76 051 489
Lakeland Max 13.1 0.53 1.14 1.68
Dec02 Mean 3.1 -0.08 -0.06 -0.09
RMS 4.9 0.29 029 077
Min 1.7 20:397 57204915147
NASA Max 12.3 0.71 0.39 1.15
Stennis Mean 49 0.02 0.00 -0.13
RMS 5.2 0.19 017 0.44
Min 0.0 0533.5 0,33: 20 29
Japan Max 7.8 0.04 0.23 0.77
Feb03 Mean 2.8 -0.09 -0.08 :0714
RMS 3.7 0.14 0.21 0.56
Min 0.0 3900.98. 27082
Ajax03 Max 22.7 0.56 044 0.67
0.2m GSD Mean 4.6 004 -0.01 012
RMS 5.7 0.22 024 0.31
Lakeland Min 3.0 -0.36 -0.24 -0.58
Mar04 Max 11.9 0.16 028. 071
0.3 m Mean 5.5 -0.13 20.05 0.09
GSD RMS 6.0 0.21 0.17: 033
Lakeland Min 3.0 -0.29 -0.14 -0.34
Mar04 Max 9.0 -0.04 0.42 0.21
0.2 m Mean 5.0 -0.14 0.05 -0.10
GSD RMS 5.1 0.17 022 0.22
Table 2: Statistics of Checkpoint Residuals for Different DSS
Test Flights - 55 mm lens
-
e
|
2
eo
e
o
a
S
RMS of Checkpoint Residuals (m)
e e
eo >
1800 m AGL
ORMS_X HRMS_Y ORMS_Z
1200 m AGL
Figure 9: Statistics of Checkpoint Residuals for Lakeland Jan04
Test Flight - 35 mm Lens - Two Flight Altitudes
Note that in Table 2, the first three flights have a common GSD
of 0.3 m, while the Ajax 03 flight and the Lakeland Mar04
flights have a 0.2 m GSD which explains the rather higher
accuracy, especially in the elevation. The first column in the
statistics in Table 2 shows the remaining y-parallax, which is
well within one pixel, which indicates the great capability of the
DSS data for stereo mapping. Figure 8 shows the RMS of
checkpoint residuals for the test flights flown with a 55 mm
lens. Note the different GSD imposed on the Ajax03 flight in
Figure 8. Figure 9 shows the RMS of checkpoint residuals for
the Lakeland Jan04 flight flown using a 35 mm lens for the two
different flight altitudes. It is clear from Figure 8 and Figure 9
that the accuracy of the checkpoints is within '^ to % of the
GSD in the horizontal components and around 1 to 2 times of
the GSD in the elevation.
S. INTEGRATED SENSOR ORIENTATION
THE DSS FOR LARGE SCALE MAPPING
USING
To evaluate the performance of the DSS for large scale
mapping applications, the DSS has been flown by PASCO
Corporation over Tovonaka city in Japan in February, 2003. For
details, see Tachibana et al (2004). The DSS was mounted
onboard an AS350B helicopter shown in Figure 10 which was
flown at about 300 m flight altitude, resulting in 0.05 m GSD.
At this rather large scale, the GPS data would be the dominant
source of error as far as the geometric accuracy is concerned.
Therefore, the DSS data has been processed using the
Integrated Sensor Orientation method (ISO) where an
aerotriangulation scheme is used to process all the data to come
up with the best fit. For details, see Heipke et al (2001). The
results of the test flight are described in some detail in
Tachibana et al (2004).
Figure 10: AS350B Helicopter Used for PASCO Toyonaka
0.05m GSD Test Flight in Japan
