rallaxes 
can be 
ements, 
p of the 
er ele- 
lation 
as the 
the dy 
(41) 
ror of 
> since 
ciple, 
e, can 
serva- 
ments 
domi- 
lected 
deter- 
er of 
under 
r and 
vhich 
  
    
  
   
  
  
  
  
  
   
      
  
  
  
  
   
  
  
   
  
     
   
     
SYMPOSIUM—NON-TOPOGRAPHIC PHOTOGRAMMETRY 105 
can be neglected in practical work. Consequently, test measurements are neces- 
sary for photogrammetric equipment and procedures in order to determine the 
standard error of the observations. T'he more carefully this work 1s done the less 
is the approximate character of the formulas for the error propagation. 
1231 PRACTICAL APPLICATION 
The formulas (34)-(37) are applied to the data in Table 1 and points 1, 2, 3 
and 4 are chosen for the determination of corrections. 
The obtained corrections are: 
dbx = — 27.3 mm. 
dbys = — 15.3 mm. 
db, = 0.000422 radians = 1.4 
dy, = 58 mm. 
After applying the corrections to the discrepancies of the rest of the control 
points, there is found the mean square value of the residuals in the remaining 
control points of 20 mm. A considerable improvement of the y-determination 
has thus been obtained by the used simplified method for determination of 
corrections to the elements of orientation. Due to the strong correlation between 
the elements of orientation, the corrections can vary rather much for different 
combinations of control points. 
In many cases in practice it may of course be difficult to place control points 
in accordance with Figure 10. But particularly in non-topographic applications 
a similar arrangement often ought to be possible. An easy check of the accuracy 
of the elements of orientation is in such cases possible, and corrections can 
obviously be determined if necessary. 
SUMMARY 
This paper contains a concentrated demonstration of the theory of errors 
for the terrestrial stereophotogrammetric normal case. Particularly the con- 
nection between the errors of the elements of orientation and the errors of the 
distances y from the base to arbitrary points has been treated, since the dis- 
tance determination normally is the most critical. The error propagation from 
the elements of orientation has been demonstrated before and after an adjust- 
ment of discrepancies in control points. In a practical example the theory has 
been applied to the determination of the corrections of the elements and the 
error propagation. 
Finally, a simplified procedure for the determination of the corrections from 
discrepancies in suitably located control points has been demonstrated. Such 
a procedure is of particular interest for many non-topographic applications of 
photogrammetry, for instance, culture-historical purposes. In such cases the 
final plotting of the pictures can be expected to be performed several genera- 
tions after the photography. The operator will surely appreciate an easy way 
to correct the errors which were made in connection with the photography, 
particularly if the feature is no longer available for a direct check. 
For testing of instruments and operators, the principles, demonstrated above, 
may have particular use. 
REFERENCES 
; Finsterwalder, R.: Photogrammetrie, Walter De Gruyter and Company. Bedin, 1052. (German) 
; v Gruber, 0: Photogrammetry—Collected Lectures and Essays, American Photographic Pub- 
lishing Company, Boston, 1942. 
DANN 
  
  
  
  
  
  
  
  
  
  
  
  
  
a 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
n 
AN : 
  
  
  
At 
EEE E RI D ENSURE 
PE SERRE IE TE 
ae 
  
  
  
  
  
  
te reel