522.2 Measurement 
The measurement of heights should be done with a carefully adjusted precision 
stereo-plotter. The ice models are best orientated numerically, the vertical para). 
laxes being measured stereoscopically. At every one of the 95 intersections the 
height is measured at four adjacent points, one in each “quadrant”. The mean 
figure is considered as the height value of the point of intersection. 
522.3 Computation of ‘the distortion” for each circle from the measured 
height values 
The problem is to compute the 9 dr-values from 95 height observations in a 
stereo model. This is a problem of adjustment and the solution is here made ac- 
cording to the method of least squares. 
The correction equations can generally be written 
ds LEA 
V - H,= gr zum EU mn 
r l > Jj S 
Vas] + r 
b 2 
b V! ; 
V ( + 2 + 
The origin of coordinates is choosen in the centre of the model and the x-axis 
is in the base direction. 
dr, — dh, 
V, correction to the observation 
H, the measured height value 
dr; the distortion in the left picture 
dr, 9 99 33 99 right ER 
b the base 
C the focal length of the camera 
dh, translation in height 
In this case the influence on the height values from the reciprocal and absolute 
orientations has been neglected. The correction equations can easily be completed 
with terms which content the influence from the orientation elements, but prac 
tical tests have shown that the reciprocal orientation has very little influence 
of the result if the distortion values are small. 
If the mean of all height observations is computed and this value is subtracted 
from the individual height observations we can eliminate dh, and if then the mean 
of the height observations (H,) in symmetrically situated test points (Hye, Hi 
Haa, Ha») is computed the influence of the absolute orientation is eliminated. 
The coordinates of the test points are tabulated in table 8. 
The correction equations are shown in sheme 1. 
The normal equations are solved according to the method of Rubin-Cholesky 
in sheme 2. 
28 
    
     
  
  
   
   
   
    
  
  
    
     
   
      
  
     
   
   
  
  
    
   
     
  
   
    
Table 8. Coord 
UN ss 
Point 
———————— 
06 m, 60 m 
07a, 07 b, 70 a, 
08a, 08 b, 80a, 
09 a, 09 b, 90 a, ' 
15 m, 51 m 
16a, 16 b, 61a, ‘ 
24 m, 42 m 
25a, 25 b, 52a, | 
26a, 26 b, 62a, ( 
835m 
94a, 81b, 43 a, 
95a, 355, 583, 7 
36 a, 36 b, 68 a, t 
14a, 44b 
  
The N-factors i 
C 
Ny =1 (2 Hi; 
C 
Nez = CH 
€ 
Ny =, (2H, 
C 
Ny, = (2 H,,