ESI 
INLINE 
ie 
EEE. 
  
  
  
  
146 THE CANADIAN SURVEYOR May, 1960 
RrsULTs oF STEREO GRID MEASUREMENTS 
According to experiment specifications, the participants submitted stereometric 
grid readings which could be used to determine the precision of the plotting equip- 
ment. Also an attempt was made to correct spot elevations for instrumental errors 
shown by grid measurements, but no consistent results were obtained. It appeared 
that the quality of the plotting equipment represented in this project was generally 
very good and that the uneveness of the grid models is, with few exceptions, well 
within the order of accuracy characteristic of photogrammetric operation as a whole. 
It was suspected also that in some cases grid measurements were not sufficiently 
reliable to provide information on the accuracy of the plotter concerned. Therefore 
grid measurements were not included in correcting the final results, but they are listed 
in Table III here as interesting information. 
TABLE III—REsULTS oF STEREO GRID MEASUREMENTS 
  
  
ma from grid measurements 
  
1 
Number of | expressed in m on the ground m, expressed in horizontal 
instruments assuming 1:50,000 photography | parallaxes in photo scale 1:50,000 
70 87 NAT Te wm a a. m 
2 x C-8 | +0.72 m | +9 y 
9 X A-8 | +0.38 m +5 u 
8 X Kelsh | +0.52 m | +6.5 u 
  
MAGNITUDE OF RANDOM ERRORS IN SPOT 
ELEVATIONS 
  
     
It is quite obvious that the results listed in 
Diagonal through central Table II are not free from some remaining 
point of the model. systematic errors. In particular, they are not 
free from the inevitable errors in absolute 
orientation and errors resulting from defoma- 
tion of the optical model caused by environ- 
mental, instrumental, and human factors. For 
deriving final figures characteristic for purely 
accidental errors, which with some restrictions 
could be considered as a measure of the inner 
accuracy of the plotters used, it was necessary 
to free the results from detectable systematic 
errors. This was done by projecting all the 
elevation points on a diagonal in such a way 
that their distance from the central point on 
the diagonal, taken as the central point of the 
model, would remain unchanged (Fig. 4). 
  
At these points the crude elevation errors 
were plotted perpendicular to the diagonal 
(Fig. 5). In this presentation much of the 
systematic deformation of models was visible 
  
  
  
at first glance and could be corrected graphi- 
Fic. 4 cally by applying either a parabolic or a linear