the
the
| the
er ef
sen
. for
ove
IEE’S
and
Vset
veen
turn
best
(ters
this
best
than
ntial
nent
ible,
1 ob-
1clu-
sults
DG
ge if
Ising
ibra-
aron
d for
cipal
er of
ight,
Ore).
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part Bl. Istanbul 2004
point and boresight parameters could be accurately determined
with blocks consisting of a single bi-directional flight line and
no GCPs. With the I-block the determination of some image
deformation parameters was not possible, so in accurate
applications the use of the cross-shaped block is recommended.
These results prove theoretical results of Honkavaara (2003).
Single GCP improves the reliability, but does not enable
accurate determination of the principal distance or datum in the
single scale calibration. In accurate applications about 10
GCPs are preferable, as suggested by Honkavaara (2003). Even
a single GCP is advantageous, if the unknowns to be estimated
are significantly larger than the obtainable standard deviations.
The standard deviations of the unknowns have been
extensively used as quality indicators in this study. As well
known, there are many limitations for the utilisation of these
numbers (systematic errors, improper weighting, ignored
correlations etc.), and they cannot be used as a sole quality
indicator. It is recommendable to collect standard deviation
data systematically, because this is valuable information for the
quality control procedures.
6. CONCLUSIONS
In this article results of analysis of 10 calibration blocks of sca-
les 1:8000 and 1:16000 were given. The block structures were
a comprehensive block consisting of 8 flight lines and having
12 GCPs, a single bi-directional flight line having one double
GCP and a single bi-directional flight line without GCPs.
Calibrations were made in single scales.
The most significant calibration parameters were interior ori-
entations and boresight. Also most of the physical and Ebner’s
mathematical deformation parameters were significant. Tests
are still needed to conclude whether or not these additional
image deformation parameters improve significantly the quality
of DG; with the examined optics it appeared that rather the
quality of the orientation observations was the bottleneck.
These results do not verify what are the best additional para-
meters to be used. Interior orientation parameters have a clear
physical meaning, thus their use is well founded. Full sets of
radial and tangential distortion parameters are difficult because
they correlate with other parameters. It is important, that the
treatment of the deformation parameters and the analysis of the
correlations and accuracy are efficiently implemented to the
commercial block adjustment software.
Principal point and boresight parameters could be determined
accurately enough with a block having a single bi-directional
flight line and no GCPs. The use of one double GCP was
advantageous only if >20 cm datum corrections or >20 um
principal distance corrections were necessary.
REFERENCES
Baron, A., W. Kornus, J. Talaya, 2003. ICC Experiences on
Inertial/GPS Sensor Orientation. In: Proc. of Workshop:
Theory, Technology and Realities of Inertial/GPS/Sensor Ori-
entation, ISPRS WG I/5, Barcelona 2003, on CD-ROM, 5 p.
171
Brown, D.C. 1976. The Bundle Adjustment — Progress and
Prospects- International Archive of Photogrammetry, Com. III,
Helsinki, Finland.
Burman, H, 2000. Calibration and Orientation of Airborne
Image and Laser Scanner Data Using GPS and INS.
Dissertation, Kungl Tekniska Hogskolan.
Cramer, M, 2001. Performance of GPS/Inertial Solutions in
Photogrammetry. In: Fritsch/Spiller (eds.): Photogrammetric
Week ‘01, Wichmann Verlag, Heidelberg, Germany, p. 29-62.
Cramer, M, Stallman, D., 2002. System Calibration for Direct
Georeferencing. In: International Archives of the
Photogrammetry, Remote Sensing and Spatial Information
Sciences. Vol. 34, Part 3A. pp. 79-84.
Ebner, H., 1976. Self Calibrating Block Adjustment. In:
International Archives of Photogrammetry, Vol. 21, Comm IIL
Fraser, C., 1997. Digital Camera Self-calibration. ISPRS
Journal of Photogrammetry and Demote Sensing. 1997; 52(4):
149-159.
Honkavaara, E., llves, R., Jaakkola, J., 2003. Practical Results
of GPS/IMU/camera System Calibration. In: Proc. of
Workshop: Theory, Technology and Realities of Inertial/GPS/
Sensor Orientation, ISPRS WG I/5, Barcelona, CD-ROM, 10 p.
Honkavaara, E., 2003. Calibration Field Structures for
GPS/IMU/Camera-system Calibration. The Photogrammetric
Journal of Finland, Vol. 18, No. 2, pp.3-15.
Jacobsen, K., 1982. Attempt at Obtaining the Best Possible
Accuracy in Bundle Block Adjustment. Phtogrammetria,
37(1982) 219-235.
Jacobsen, K. 2003. Issues and Method for In-Flight and On-
Orbit Calibration, Workshop on Radiometric and Geometric
Calibration, Gulfport, 2003, 11 p.
Kilpelä, E., 1981. Compnensation of Systematic Errors of Ima-
ge and Model Coordinates. Photogrammetria, 37(1981) 15-44.
Schroth, R., 2003. Direct Georeferencing in Practical
Applications. In: Proceedings of Workshop: Theory,
Technology and Realities of Inertial/GPS/Sensor Orientation,
ISPRS WG 1/5, Barcelona 2003, on CD-ROM, 5 p.
Tempelmann, U., Hinsken, L., Recke, U., 2003. ADS40 Calib-
ration & Verification Process. In: Proc. of Workshop: Theory,
Technology and Realities of Inertial/GPS/Sensor Orientation,
ISPRS WG 1/5, Barcelona 2003, on CD-ROM, 6 p.
Wegmann, H., 2002. Image Orientation by Combined (A)AT
with GPS and IMU. In: Proceedings of ISPRS Commission I
Mid-term Symposium, Nov. 2002, Denver, Colorado, USA.
ACKNOWLEDGEMENTS
I am grateful to Risto Ilves and Juha Vilhomaa from the Natio-
nal Land Survey of Finland for providing me extensive data for
the empirical investigation. I also thank Eero Ahokas from the
FGI for carefully building and maintaining the calibration
fields that were used in this study.
|
i
i
|
