42 
concern aspects of the hardware - all calibration 
procedures on the S9AP already run 
automatically. This concerns more the quality 
control aspects of obtaining good points. 
Two methodologies used in this investigation 
would be difficult to implement in an automated 
system: adjustment of the video digitisation gain 
and offset, and the selection of best matching 
parameters. 
The former may always be a problem, due mainly 
to the density range in aerial photography. 
However in this case it was necessary due mainly 
to the poor contrast exhibited by many of the 
signalised points used. In a semi-automatic 
procedure the points to be matched will in general 
be natural tie-points instead of signalised points. 
Images in the appropriate positions in the model 
can be acquired (eg at the von Gruber points) and 
the tie-points selected automatically based on what 
image signal there is. The Digital Comparator 
Correlator System (DCCS) from Helava Associates 
Inc employs techniques along these lines in its 
procedures for semi-automatic triangulation 
(Helava, 1988). 
The need to select best matching parameters was 
also a direct consequence of the signalisation - 
badly defined points are difficult to match with 
template matching. This would also to some extent 
be relieved by the use of natural tie-points: in 
general, good signal content equates to suitability 
for matching. If a constant parameter set is used, it 
may be easier to pick out any failures that do occur, 
whereupon alternative points could then be 
selected on-line. It may still be necessary to match 
signalised points, in which case a higher quality 
signalisation is required, or the development of 
supporting or alternative matching methods. 
5. Concluding Remarks 
The primary aim of the investigation was to use the 
results of the bundle adjustments as a means to test 
the quality of the calibration procedures developed 
for the CCD cameras on the S9AP. The accuracy 
level achieved with the digital measurements was 
slightly lower than that for the manual 
measurements. With the test data used, it is 
difficult to separate out the exact source of this loss 
of accuracy: some certainly lies in the image 
matching techniques used, whilst small instabilities 
in the calibration may have contributed also. The 
conclusion for the calibration is that it is at an 
adequate level for the actual purpose for which it 
has been developed - DTM generation by image 
matching. 
The secondary aim was an investigation of aerial 
triangulation with point determination by image 
matching. The accuracy level is comparable with 
manual observations, but a more thorough 
investigation is required to determine the influence 
of poor matching on the result. Aspects of semi 
automation of the procedure have been addressed, 
where the use of natural tie-points would provide 
some additional advantages. If signalised points 
are used, their imaging quality must be high. 
6. References 
ACKERMANN, F. and SCHNEIDER, W., 1986. High 
Precision Aerial Triangulation with Point 
Transfer by Digital Image Correlation. 
International Archives of Photogrammetry and 
Remote Sensing, 26(3/1): 18-27. 
BALTSAVIAS, E.P., 1988. Hierarchical Multiphoto 
Matching and DTM Generation. International 
Archives of Photo grammetry and Remote 
Sensing, 27(B 1 l/III): 476-487. 
BETHEL, J., 1986. The DSR11 Image Correlator. 
Proc. ASPRS Annual Convention (Washington 
DC), (4): 44-49. 
DAHLER, J., 1987. Problems in Digital Image 
Acquisition with CCD Cameras. Proc. ISPRS 
Inter-Commission Conference on Fast 
Processing of Photo gramme trie Data (Interlaken, 
DI Cf 
}j S c 
Switzerland), pp 48-59. 
1 he 
GRÜN, A.W., and BALTSAVIAS, E.P., 1988. 
Geometrically Constrained Multiphoto Matching. 
Photogrammetrie Engineering and Remote 
Sensing, 54(5): 633-641. 
pic 
r 0 i: 
The 
GRÜN, A.W., and RUNGE, A., 1987. Projekt 
«Heinzenberg» - Präzisionsphotogrammetrie im 
Gebirge. Vermessung, Photogrammetrie, 
Kulturtechnik, 87(7): 271-276. 
GÜLCH, E., 1984. Geometrie Calibration of Two 
CCD Cameras used for Digital Image Correlation 
on the Planicomp C 100. International Archives 
of Photo grammetry and Remote Sensing, 
25(A3a): 363-372. 
Helava, U., 1988. Digital Comparator Correlator 
System. International Archives of 
Photo grammetry and Remote Sensing, 26(B2): 
160-170. 
PERTL, A., 1984. Digital Image Correlation with 
the Analytical Plotter Planicomp C 100. 
International Archives of Photo grammetry and 
Remote Sensing, 25(A3b): 874-882. 
SCHNEEBERGER, R., and BÜRGERMEISTER, W., 
1987. The New Wild System-9 Analytical 
Stereoplotter Workstation S9AP. Proc. ASPRS 
Annual Convention (Baltimore), (2): 180-189. 
If