Phase III: Block II (two additional cross strips), post-
triangulation.
This paper summarises the outcomes of the testing. Based on
the findings of the research, recommendations are made as to
the most suitable block configuration to be selected when using
ADS40 data of this kind.
2. TEST DATA AND SOFTWARE
2.1 Test data
The dataset made available for the study consisted of six strips
of ADS40 imagery with the characteristics shown in Table 1.
ADS40 imagery Six strips with GPS/IMU data
Coverage 8 x 8 km
Location Waldkirch, Switzerland
No. of GCPs 30
Focal length 62.5 mm
Flying height 2000 m
Pixel size 6.5um
GSD 200 mm
Table 1. Test data characteristics
2.2 Software
Software made available for the purpose of conducting the
study comprised:
eADS40 ground processing software, GPro (v. 2.18)
eSOCET SET photogrammetric software suite (v. 4.3.1)
eORIMA bundle adjustment software (v. 6.0)
3. PHASE I: BLOCK I, PRE-TRIANGULATION
3.1 GPS/IMU Processing and image rectification
The raw ADS40 images were first resampled to remove the in-
fluences of the aircraft movement during image acquisition.
This was performed using position and attitude data from a
variant of the Applanix Position and Orientation System (POS)
developed to meet the requirements of the ADS40 sensor (San-
dau et al., 2000). Data was collected synchronously with image
acquisition and processed relative to a GPS base station. This
stage of the pre-processing was performed in-house by Leica
Geosystems, investigation of this aspect of the processing flow-
line being beyond the scope of this research study. Rectification
of the ADS40 imagery was performed, in order to produce ste-
reo-viewable images, using GPro. The first stage of the study
was concerned with assessing the three possibilities for stereo
viewing of the rectified panchromatic images, i.e. F/N, F/B and
N/B scene configurations, as described by Templemann et al.
(2000).
3.2 Assessment of different stereo configurations
The three different stereoscopic combinations were assessed us-
ing stereo measurements of 18 independent GPS coordinated
check points. Figure 2 shows the check points that were avail-
able for measurement using the SOCET SET photogrammetric
software suite.
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part Bl. Istanbul 2004
WS SE s
ECO
* u.c. tle. uU 4.
Figure 2. Location of the independent check points available for
stereo measurement
The final results of stereo measurements for each combination
are presented in Table 2. Each different measurement combina-
tion produced different results, partly as a consequence of the
different stereo angles (Figure 1). A reasonable RMS of ap-
proximately two pixels was present in the horizontal compo-
nents (X, Y), however the vertical component (Z) was greater
than six pixels in each combination, which does not satisfy the
requirements of photogrammetric measurement. The three dif-
ferent possible stereo configurations were therefore considered
in the subsequent phases of this study to find the impact of each
combination on photogrammetric application.
Scene RMS X RMS Y RMS Z
combination (m) (m) (m)
F/N 0.323 0.267 1.316
F/B 0.358 0.247 1.263
N/B 0.431 0.204 1.310
Table 2. RMS of Phase I stereo measurements
4. PHASE II: BLOCK I, POST-TRIANGULATION
4.4 Block I triangulation
The quality of the georectified imagery may be improved
through triangulation in order to achieve a higher level of preci-
sion and accuracy. The four parallel strips of rectified imagery
were therefore triangulated using ORIMA; first of all without
and subsequently with three different configurations of GCPs.
Four different combinations (F/N, F/B, N/B and F/N/B) of the
three panchromatic scenes were triangulated.
4.0 Ground control point configurations
The GCPs configurations, selected as subsets from the dataset
of 30 GPS coordinated ground points provided, comprised zero,
four, nine and 12 GCPs (Appendix A, Figures A1 to A4).
4.3 Results from Block I adjustments
The Combined Adjustment Program for Aerial Triangulation
(CAP-A), which is fully integrated into the working environ-
ment of ORIMA, was used to perform a bundle adjustment for
the 16 block combinations. Accordingly CAP-A outputs were
checked to eliminate blunders. Errors affecting o, and each
ORIMA block output were analysed. The final adjustment pre-
cisions for each triangulation are listed in Table 3.
Interna
comb
UU € on
F/
It is no
slight e
combir
duced :
44 €
The inc
ured ir
combir
three s
mary c
Appenc
process
uremen
measur
stereo |
It is cle
the low
importi
the acc
importi
is ther:
what is
compoi
measur
figurati
curacy,
graphs.
but furt
It can |
chroma
The F/
measur
the bes:
S.P
Si F
In Phas
so that
followe
of impc
ferent ‘
tions as
52 €
The GC
dataset,
cated as