A general view of the object and the spatial 
control point configuration from one of the 
camera stations is shown in /1/. Out of 
Table 1 we can see that most of the control 
points are imaged in all slides. 
  
slide control points 
  
1 12. 15, 44, 45, 16, 18, 19, 21, 
22, 23, 24, 25, 26, 27 
2 12, 13, 14, 15, 16, 17, 18, 19, 21, 
22, 23, 25, 25, 26, 27 
5 ho, 12, 15, 14, 15, 16, 17, 18, 19, 2, 
  
  
  
  
25, 25, 26, 27 
4^ ho, 12, 15, 15, 15, 15, 17, 13, 19, 21, 
22, 25. 25, 26, 27 
Table 1 
The computations were performed with a 
FORTRAN IV program on the Cyber 17H of 
Stuttgart University. This program allows 
apart from the case of an orientation to be 
used as a calibration program without regard 
to lens distortion or with one (ease I), 
three (case II) or five (case III) distortion 
parameters. 
The data were processed with these four 
calibration versions for three control point 
configurations. In Table 2 and Table 5 all 
points are considered as control points 
(configuration 1). In Table 4 we consider all 
points except 12, 14, 19, 21 as control 
points (configuration 2). In addition the 
points 25 and 26 are not considered as 
control points (configuration 3). These 
results are listed in the lower part of 
Table 4, 
Some of the results obtained with the 
calibration program without regard to lens 
distortion are given in the upper part of 
Table 2. The differences DX, DY, DZ between 
the given object space coordinates and the 
computed coordinates are remarkable. We find 
the greatest differences in the DY-column, a 
reasonable fact as the camera axis points in 
the direction of the y-axis. The next column 
RD eontains the resultants of the differences 
in the x-z-plane in the photo scale. The 
ratios between the resulting differences in 
the x-z-plane and the object distances 
(in per mille) are given in column R%o. 
In the last column Y%othe ratios between 
DY and the object distances are listed. 
The lower part of Table 2 shows the cor- 
responding results obtained with the 
calibration program with distortion case I. 
Comparing with the upper part we can see the 
improve of the accuracy. For our example we 
cannot realize a gain of accuracy by using 
the distortion cases II and III (upper and 
lower part of Table 3) in comparison to 
case I. That is why we consider in Table 4 
only the distortion case I for the control 
point configurations 2 and 3. 
6. CONCLUSIONS 
The practical test of the described on-the- 
job calibration method shows at this first 
example that the use of non-metric cameras 
in industrial photogrammetry can lead to 
remarkably good results. Further in- 
vestigations concerning the distortion model, 
the configuration of control points and 
the camera dispositions are necessary to 
improve the accuracy such that non-metric 
cameras might be used unconditionally in 
non-topographic applications. 
REFERENCES 
/1/ Altan, M.0.; Bopp, H.; Krauss, H.: Some 
Accuracy Aspects of On- The-Job Cali- 
brations Shown at the Example of a 
Precision Survey. Paper, presented at the 
Inter-Congress Symposium, Commission V, 
Stockholm, 1978. 
/2/ Bopp, H.; Krauss, H.: A Simple and 
Rapidly Converging Orientation and 
Calibration Method for Non-Topographic 
Applications. Paper, presented at the 
ASP Fall Technical Meeting, Little Rock, 
Arkansas, 1977. 
/5/ Bopp, H.; Krauss, H; Preuss, H.D.: 
Photogrammetric Control Survey of a 
Large Cooling Tower. Paper, presented 
at the ASP Fall Technical Meeting, 
Little Rock, Arkansas, 1977. 
/4/ Kenefick, J.F.: Ultra-Precise Analytics, 
Phot. Eng. (1971), P. 1167 - 1187. 
/5/ Marzan, G.T.; Karara, H.M.: A Computer 
Program for Direct Linear Transformation 
Solution of the Collinearity Condition 
and Some Applications of it. Proceedings 
of the ASP Symposium on Close-Range 
Photogrammetric Systems, Champaign, 
Illinois, 1975.