5 OA VV eed 
| (VC) WwW HV) V? 
el Ey de LU) (€? 1 
ct DS V OD 
an approximative adjustment. Standard errors a priori for 
these coordinates vary between 2 mm and 5 mm, according 
to measuring accuracy. 12 photos are used (3 photos/strip 
with Ge 60% longitudinal overlap; black diamonds in 
Fig.1). 
2) 22 "bundles" from terrestrial observations of horizontal and 
vertical directions are included in the adjustment. Conside- 
ring pointing errors for shorter distances (« 100 m), stan- 
dard errors a priori are variable. Typical standard errors 
are" assigned to slant and horizontal ‘ranges as well as to 
levelled elevation differences. The calibration range is 
tied to-the first order net,“ which is considered errorless, 
thus avoiding rank deficiency. 
Number of photos corresponds to 1). 
3) All terrestrial observations are part of the adjustment, 
Corresponding ito -2). 
Number of photos is 28, positions marked by black and white 
diamonds “in Fíg.1. 
Tab.1 contains estimates of the parameters of interior orienta- 
tion. and their standard ‘deviations: a’ posteriori. Distortion 
parameters lead to differences of photo coordinates 
dx = x (Rqr? + Rar) + B,(Cy? + 3x2) + 2Boxy 
2 2 2) 
dy VAREL Ror*) + 2B xyo+ ‘Bp (%% 5+ 53y 
The full parameter set by BROWN 1976 may be completed, which 
shall be done in near future. 
Results may be commented as follows: 
1) Sigma naught is relatively large in all cases, which may be 
due to adverse circumstances (see above). Under fair condi- 
tions (better illumination), sigma naught should at least 
drop to appr. + 3 micrometers, cutting all other standard 
deviations to 75 % or less. This would favourably correspond 
to simulated calibration results (KUPFER 1986). 
2) Parameters from variations 1) through 3) are almost invari- 
ant. This shows that redundancy from measurements on 12 
photos is large enough to get proper results. 
3) Estimated calibrated focal length corresponds very well to 
that or Laboratory calibration in this case. Decentering 
distortion includes in all variations image deformation from 
film etc. This is correlated to the amounts of coordinates 
of the point of best symmetry (PPS). Despite this their 
variations from one adjustment to the other are very small. 
4) Algebraic correlation of parameters corresponds to expecta- 
tions, ie. g. 2foricalibrated: focal Length and< fiying height 
>0.9 
33