surface under condition of the least square adjust- 
ment. In the adjustment procedure the deviations 
Vx and Vy were minimalysed simultaneously for all 
9 crosses,separately for each image. The results of 
analysis of both images using 3 investigated methods 
are shown in the tab.1. 
The deviations are not big, in spite of the fact that 
the corrections for objective distortion were not 
taken into account. The mean errors mx and my 
do not exceed 0.04 pixel, what is very promising. 
Comparison of the mean errors resulting from each 
of 3 analyzed methods does not show big differences. 
In conclusion we can exclude the use of the most 
troublesome method of EV% = min parabolic ap- 
proximation, and consider for further use the method 
of linear interpolation or method of ZV“ = min 
parabolic interpolation. The further practical ex- 
ploitation of the selected methods will allow for the 
final valuation in which not only accuracy but also 
the reliability and time consumption must be con- 
sidered as an important factors. 
Tab.1. 
Residual deviations after projective transformation of 9 points test coordinates onto the digital image 
coordinate system. ( [up = 0.001 of a pixel]). 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
image,,1" image „2” 
image „1” covers full image „2” is in the scale 2.4 times 
CCD frame surface smaller then image „1” 
1 pixel = 0.55x0.37 mm 1 pixel = 1.32x0.89 mm 
A B C A B C 
Vx Vy Vx Vy Vx Vy Vx Vy Vx Vy Vx Vy 
[up] [up] Up] [xp] [up] [4p] 
1 0 19 2 27 4 30 12 11 -2 -3 21 -34 
2 -24 -34 -11 -40 -13 -53 -6 -34 9 -1 d 1 
3 9 27 5 27 6 36 -17 30 -10 24 -19 27 
4 8 5 1 13 1 16 -21 23 -27 21 -28 73 
5 10 -66 -14 -75 -14 -81 -6 -30 il 1 -10 -6 
6 11 36 18 35 19 38 49 -8 21 -64 43 -54 
7 12 10 6 4 10 8 12 9 17 -20 4 -23 
8 -26 31 5 28 -2 25 6 -23 5 2 20 -27 
9 0 -29 -13 -18 -11 -20 -29 22 -24 38 -26 44 
mx my mx my mx my mx my mx my mx my 
14.0 33.0 10.0 35:3 10.7 40.2 22.0 23.0 16.3 27.8 22.7 38.5 
  
  
  
  
  
  
  
  
Note: 
The A, B, C columns show the residual deviations of image coordinates achieved using various methods 
of the axis of the arms of cross automatic determination on the CCD digital image of the test: 
e A - method of the linear interpolation along the luminosity profile 
e B - method of the luminosity profile approximation by parabola under v 
min condition 
e C - method of the luminosity profile approximation by parabola under ZV^, — min condition. 
305