ON 
  
  
  
  
  
  
  
  
  
controlled experiments. First, the aerial photographs are probably affected by systematic disturbances 
in addition to those arising from the lens of the camera, which were determined by laboratory tests. 
Primarily there will be influences from the refraction and the earth’s curvature but there will also be a 
certain lack of flatness in the negatives; moreover large temperature differences may have a con- 
siderable influence. It is always desirable that aerial cameras in addition to laboratory tests be tested 
from the air over test fields according to the grid method. The y-parallax method for the determina- 
tion of systematic errors in the photographs is by comparison with the grid method an approximate 
one. 
The conditions of the relative orientation will usually not be strictly satisfied in empirical work. 
Even if it can be expected that the operators will perform the relative orientation during the tests 
very carefully, a correct adjustment according to the method of least squares is not to be assumed. 
The measured y-parallaxes will demonstrate the adjustment situation. Corrections to the final 
coordinates with respect to defective adjustment of the relative orientation are not intended to be 
computed. Therefore the mean square value of the residual y-parallaxes from measurements in 15 
points will be used as a substitute for the standard error of unit weight of the y-parallax measurements. 
Obviously this also means a certain approximation. 
If there are redundant control points, discrepancies will usually appear which in practice are 
adjusted more or less arbitrarily. The formulas for the error propagation are founded upon the 
assumption that such discrepancies are adjusted according to the method of least squares. In the 
simplified formulas for the accuracy of the final coordinates and elevations which will be used for 
the predictions of accuracy it is further assumed that the standard error of unit weight of the y- 
parallax measurements is of the same magnitude as the standard error of unit weight of the image 
(model) coordinate measurements. 
Particular attention will be paid to the question of the statistical distribution of the measured 
y-parallaxes and the residual y-parallaxes after an adjustment of the relative orientation. 
Finally, it must be emphasized that the majority of the formulas which are to be used for the error 
propagation studies relate in principle to comparatively flat ground and approximately vertical 
photographs. In partieular the photographs for the controlled experiments of Sub-commission IV:1 
were photographed with comparatively large inclinations and over terrain with very large elevation 
differences. The influence of these unfavourable factors upon the accuracy determination will be 
directly studied too. The basic formulas for the treatment of the adjustment of the relative orienta- 
tion and the error propagation will be derived for arbitrary elevation differences. 
l. GRID TESTS OF THE PLOTTING INSTRUMENTS 
1.1 Programme for the grid measurements in the individual projectors of the 
instruments 
A glass grid of high and known accuracy must be available for the measurements. Depending upon 
the setting of the base, two different combinations of the grid points will be used for the measurements. 
1.11 Base zero. 
After careful adjustment of the grid in the projector, which must be in zero position, the machine 
coordinates of the grid points according to Fig. 1 are measured in the same average enlargement as 
during the controlled tests. In every point each coordinate must be determined by at least three 
independent settings, readings and recordings. Special forms are to be used (see Appendix 1). 
1.12 Base in or out. 
Only one base setting, corresponding to a longitudinal overlap of 50 95, is assumed for the measure- 
ments. The same average enlargement as during the controlled tests is assumed. 
4 
o3 
o1! 
Fig 
not 
Fig 
of 
the 
foi 
th 
ot 
nc 
ex 
eC 
tv