28 
The following distribution is found: 
0.7 0.8 0.9 1.0 1.1 1.2 
1 8 12 6 3 2 
Assuming approximately the same standard error in the grid coordinates and in measurement with 
the comparator the standard error of measurement with the comparator becomes about 0.6 /nm. 
This figure may be regarded as the present limit of accuracy in photogrammetry. 1 ) 
Example of histogram and normal distribution curve. Residuals 
after test measurements in a stereocomparator and adjustment. 
Number 
of residuals 
s = 0 94 microns 
Fig. 2.1:1. Normal distribution hypothesis is accepted. 
5.5 < chi 2 5 = 16.9 
Tolerances for arbitrary functions of the basic measurement in the comparator can be established 
according to the principles of statistics, provided the basic errors are normally distributed on an accept 
able level, fig. 2.1:1. Similar problems were discussed in reference 2.1:4. 
References. 2.1:1. Lycken, L. E.: Test Measurements of Grids. Int. Archives of Photogrammetry, Yol. 
XII, Stockholm 1956. 
2.1:2. Hallert, B.: Test Measurements in Comparators and Tolerances for Such Instruments. 
Photograminetric Engineering March 1963. 
2.1:3. — Bestimmung der Präzision und Genauigkeit eines Stereokomparators. Schweize 
rische Zeitschrift für Vermessung, Kulturtechnik und Photogrammetrie. 
1963, Nr. 9. 
2.1:4 — Tolerances of Photogrammetric Instruments and Methods. Invited Paper. 
Comm. II. The Int. Congress for Photogrammetry Lisbon 1964. 
x ) Another stereocomparator PSK (Zeiss-Aerotopograph) has been tested in Finland with the same glass grid Nr 410. 
From Professor Dr. R. S. Halonen the following preliminary results have been received, April 25, 1964: 
V' m 
dfi cc 
10" 6 
Left plate holder (averages of 11 measurements) 
2.01 
— 9.6 
+ 34.1 
Right » » (averages of 12 measurements) 
1.95 
-20.1 
+ 11.5 
The standard deviation of one setting was found to be 0.8 um. 
The final results will be published in Communication Nr 4 from the Finnish Society of Photogrammetry.