d 
ly 
1e 
1€ 
ts 
  
  
so fundamental for the entire procedure, were obtained. Since the relative orientation of the models 
in general was empirically adjusted the mean square value of the measured y-parallaxes was used 
as a substitute for the standard error of unit weight of the y-parallax measurements. The order of 
magnitude of the mean square values of the parallaxes was found to be 8&—10 microns on the scale 
of the photographs (diapositives) for the wide angle images and about 7 microns for the normal angle 
images. (See Tables 2 to 6.) The standard error of unit weight of the parallax measurements was 
computed for all models from 15 and 9 orientation points. (See Tables 2 to 6.) In the models over 
rugged ground (Monti di Revôira) formulas for flat and for rugged ground were used for comparison. 
In this particular case no significant difference between the results from the two formula systems 
could be found. (See Table 1.) 
The theoretical relation between the standard error of unit weight of the y-parallax measurements 
and the mean square value of residual y-parallaxes after an adjustment according to the method of 
the least squares was tested practically. A very good agreement was found. The distribution of the 
errors of the fundamental operations is of particular interest for the treatment of the error propagation. 
The distribution of the y-parallaxes before and after the adjustment was studied in histograms and 
tests were performed according to ordinary statistical procedures, whether the distributions were 
normal or not, see Diagrams 1 to 4. The distributions proved to be normal with high reliability. 
y-parallaxes were also used for determinations of possible radial distortion of the reconstructed 
bundles of rays in the instruments. From the photographs of the rugged test area Monti di Revoéira 
no consistent radial distortion could be determined in this way, probably owing to the local circum- 
stances (rugged ground, comparatively large inclinations of the photographs, fluctuations of the 
refraction of the atmosphere). In other test photographs over ground with moderate elevation differen- 
ces normal results of the radial distortion determination were found, see Diagram 7. 
Some tests were also applied to the diapositives which were used for the actual test measurements 
of Sub-commissions IV:1 and 1V:3. In some cases considerable deviations from flatness were found, 
see Diagrams 8 to 15. From measurements of fiducial mark distances considerable shrinkages of some 
film negatives were found. 
From the results of the y-parallax measurements, estimations were made concerning the accuracy 
to be expected in the final coordinates and elevations under well-defined conditions. The accuracy 
was expressed as mean square values of discrepancies in the check points of the models. The predictions 
of the accuracy to be expected were also furnished with confidence intervals in order to facilitate the 
judgments of the discrepancies between the theoretically predicted and the practically determined 
accuracy of the final results. Finally, some previous test results were summarized. 
In Appendix 1 complete formula systems and detailed results of performed computations are 
demonstrated. In Appendix 2 the results of an international investigation of existing specifications 
for aerial photogrammetry are summarized. 
REFERENCES 
1. HArLERT, B.: International Comparative Tests. — A Suggestion. Int. Archives of Photogrammetry. Vol. XII 1956. 
2. Uber die Genauigkeit der Luftphotogrammetrie. Transact. of the Royal Institute of Technology, Stockholm, 
1956, No. 99. 
3. A New Method for the Determination of the Distortion and the Inner Orientation of Cameras and Pro- 
jeetors. Photogrammetria 1954— 55, No. 3. 
4. HarLerT, B., REENLUND E.: Suggestion for a Standardized Test of the Projection System in Photogrammetric 
Plotting Instruments. Svensk Lantmiteritidskrift 1956, Congress Number. 
HarrERT, B.: The Grid Method and the y-parallax Method for the Determination of Systematic Disturbances 
in Aerial and Terrestrial Photographs. Svensk Lantmäteritidskrift 1956, Congress Number. 
6. Results of y-parallax Measurements in a Wild Stereocomparator and Some Additional Notes to the y- 
Parallax Method for the Determination of some Types of Systematic Disturbances in Aerial Photographs. 
Photogrammetria 1958— 1959, No. 2. 
7. CRAMÉR, H.: Mathematical Methods of Statisties. Princeton and Uppsala 1945— 1946. 
8. HarrERT, B.: Praktische Untersuchungen über die Verwendung der Methode der kleinsten Quadrate in der 
St 
Photogrammetrie. Zeitschrift für Vermessungswesen 1959, No. 1. 
27