21 
1.62 Oland Field Tests 
A series of tests of aerial photographs have been made mainly to determine the standard error of unit 
weight of image coordinates under operational conditions. The principles of the tests are identical with 
those described in reference 1.61:1. The flying altitudes chosen for the cameras to be tested were 1200 
m above ground for wide angle cameras and 2600 m and 900 m for the superwide angle camera. In the 
spring of 1963 five wide angle cameras and one superwide angle camera were used, and in the fall of 
1963 three wide angle cameras which were used earlier were retested. 
In all cases approximately vertical photographs were taken above the center point of the test field. 
In all wide angle photographs, treated here, Kodak Aerographic Super XX film on acetate base was 
used. The image coordinates were measured in the Wild StK 824 stereocomparator by different operators. 
The computations and adjustments were made in the FACIT electronic computer after programming 
by K. Torlegdrd and B. Adolfsson, MSE. The following data were determined through the adjustment 
operations: 
Radial distortion of the photographs in two orthogonal directions, including a determination of affine 
deformations 
Standard error of unit weight of image coordinates for circles with different radii 
Residual image coordinate errors after the adjustment in a specified circle and after correction for 
radial distortion and affine deformation 
The results are presented here with diagrams for each photograph in the Appendix. In each case the 
distribution of the residuals is shown and tested for normality. Almost always the residuals were nor 
mally distributed on the 5% or at least 1% level. N—S means flight direction North—South and 
E—W, East—West. 
It is noted that: 
There is considerable affine deformation in all photographs, causing differing radial distortion curves 
in the x'~ and the y'-directions; there are considerable correlations between residual coordinate errors in 
the same photograph, caused probably by ’’regional” shrinkage effects. See section 1.63 and fig. 1.63:1. 
There is overall a rather good general agreement between the radial distortion curves of the lens and of 
the photographs (averages between the two affine distortion curves) but a minor systematic difference 
is found at a radius of about 50 mm where the radial distortion of the photographs is somewhat greater 
than that of the lens. 
The Aviogon 29 and 41 cameras were used for tests in the spring and fall of 1963. A negative radial 
distortion is more pronounced in the fall photographs at a radius of about 100 mm than in the spring 
photographs. See fig. 1.62:7 a in the Appendix. Identical magazines were used spring and fall. Atmos 
pheric refraction variations and possible deformations of the supporting plates may have had some 
influence. 
The tests confirm a considerable variation in the weights of image coordinates with change of radial 
distance from the principal point. See fig. 1.62:9 in the Appendix. 
In fig. 1.62:10 in the Appendix the results of the superwide angle photographs tests are shown 1 . 
As the signals on the ground could not be identified at a distance greater than about 100 mm from the 
principal point the radial distortion curve could be determined to this radial distance only. The agree 
ment between the radial distortion of the lens and of the photographs has been good so far. For the 
photograph from a flying altitude of 2600 m the standard error of unit weight is on the average about 
6 um and much greater in the photographs taken from 900 m. A pronounced increase with the radial 
distance is found in the latter case and is indicated in the former. The residuals are normally distributed 
in the former case on the five percent level but not at all in the latter case. 
1.63 Some Correlation Tests of Image Coordinate Residuals 
After the adjustment of the discrepancies in points on one circle the residual errors of the image 
coordinates are always more or less correlated. Most of the correlation results from the effects of shrink 
age and should therefore be considered a physical correlation. There is also an algebraical correlation 
1 These photographs were taken on Gevaert film.