it 
reduction of non-measurement error sources should 
be investigated. 
8. Conclusions 
This study shows that digital image processing can 
improve both the precision and the productivity of 
the close-range photogrammetry. In order to 
optimize these two aspects it is necessary to 
improve some parts of the digital photogrammetric 
process. 
Looking at productivity the time needed for 
digitilisation (6 hours per negative) should be 
reduced. À new version of the ASTROSCAN is 
being developed needing about half the time. If the 
need for faster digitalization arises using another 
digitizer should be considered. A CCD-camera is 
an alternative for obtaining the digital images. Then 
the problem of film deformation and therewith the 
need for reseaucrosses and their matching would be 
eliminated. 
The computational burden of the matching process 
is a second productivity aspect. The 4 hours per 
negative needed in the experiment (for an average 
negative: 121 reseaucrosses and 50 targets) using a 
80386 processor based PC. This can be reduced 
through optimization of the software and by 
utilizing a faster computer. However, these time 
apects have only importance if results are needed 
within a short timescale. It should be noted that 
digitizer and computer can operate 24 hours a day 
without manual interaction. 
As far as precision is concerned, the experiment 
showed an improvement of a factor 1.4 in variance. 
Further improvement is expected from a better film 
deformation model; also reduction of the film 
deformation itself should be aimed at. A second 
aspect influencing the precision is the camera 
model. Especially for a camera with a relatively 
short focal length (40 mm for the camera used in 
the experiment) an extension of the model with a 
few degrees of freedom is to be investigated. 
229 
Literature 
Ackermann F., 'Digitale Bildverarbeitung in der 
photogrammetrischen Stereo-auswertung 
Stuttgart, Germany, 1989 
Amdal K., Thune N., Dgrum H., ‘Correction of 
image deformations using B-spline surfaces in 
digital close-range photogrammetry' 
'Close range Photogrammetry Meets Machine 
Vision' ISPRS Symposium proceedings Zürich, 
Switzerland, Sept.1990 
Baarda W., 'Statistical Concepts in Geodesy' 
Neth. Geod. Comm., Publ. on Geodesy, New 
Series, Vol. 2, No. 4, Delft, The Netherlands, 1967 
Bihr H.P., Vogtle T., 'Digitale Bildverarbeitung, 
Wichmann Verlag, Karlsruhe, Germany, 1991 
Baltsavias E.P., Beyer H.A., Fritsch D., Lenz R.K., 
'Fundamentals of Real-Time Photogrammetry' 
Tutorial Notes, Zürich, Switzerland, Sept. 1990 
Buiten H.J., Clevers J.G.P.W. (ed.), Remote 
Sensing, theorie en toepassingen van 
landobservatie' 
Wageningen, The Netherlands, 1990 
Fórstner W., 'Prinzip und Leistunfsfáhigkeit der 
Korrelation und Zuordnung digitaler Bilder 
Stuttgart, Germany, 1989 
Fritsch D., 'Algorithms in Fast Vision Systems 
Optical 3-D Measurement Technique, Gruen/ 
Kahmen (ed.), pages 87-102, Germany, Sept.1989 
Haarlem M.P. van, Katgert P., Tritton S., 
Le Poole R.S., 'The fidelity of photographic plate 
copies' 
Mon.Not.R.A.S., Leiden, The Netherlands 1991 
Heuvel F.A. van den, 'Haalbaarheidsonderzoek 
digitale fotogrammetrie' Ing.bur. Geodelta BV, 
Rijswijk, The Netherlands, Jun. 1989