52 Some investigations into distortion of survey cameras 
521 Introduction 
Distortion in modern lenses is frequently so small that it can in practice be 
neglected for most purposes. When in 1953 RAK procured new wide-angle camera, 
it was for this reason considered unnecessary to compensate for the stated optic] 
distortion in the stereo-plotters. It soon proved, however, that there were larger 
systematic deformations in the stereoscopic models than could be expected he. 
cause of distortion. For this reason RAK started a series of experiments in 1954 
to find out the magnitude of the deformations in various parts of the model ang 
their causes. The investigations proceeded along two lines. 
First, a method has been developed for empirical determination of the magni. 
tude of the deformations. An empirical distortion curve has been computed from 
the height deformations, assuming that the deformation is due to a symmetric 
radial distortion. Parallel with this work, the optical distortion of the camem 
was determined in cooperation with the Institution for Optical Research at the 
Royal Institute of Technology. Finally, the different factors affecting the form 
of the stereoscopic model have been studied and on the basis of these data an 
attempt has been made to explain the causes of the disturbances. 
022 A method for empirical determination of the 
total *distortion^? 
Aerial photos of flat ice surfaces were used as starting material. At the end 
of April, shortly before the ice on the large lakes of Central Sweden disappears 
the ice surface is generally very rich in details and absolutely flat. This ice surface 
is then a suitable experimental field for the study of disturbances caused ina 
stereoscopic model because of optical distortion, mechanical trouble in the camera, 
curvature of the earth, refraction, etc. 
222.1 Selection of points 
If it is assumed that all disturbances of the model can be attributed to a sym 
metric radial distortion, it is practical to choose the test points on the ice surface 
so that they are on concentric circles round the two photo centres. Here nine 
9 b 
circles were chosen on each photo with radii 6 © etc. Fig. 7 shows an ice modd 
; | 
with the points of intersection of the two groups of curves. Their designations 
are stated directly in the figure. A total of 95 test points is obtained. 
26 
         
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
   
   
  
  
  
  
   
   
  
  
  
  
  
  
   
      
Fig. 7. The mod 
surface with the tw 
groups of circle. Th 
intersection points 
are used as test 
points,