must per- 
one must 
d find the 
ls case an 
we should 
ive at the 
is required 
ition equa- 
left with a 
ear in the 
vat, It has 
in a paper 
lexpectedly 
lating it to 
linear. This 
ion but for 
hieved this. 
lating by a 
it as nearly 
'et method, 
the model 
r equation. 
s not exact 
nding rays 
eviates too 
ns and we 
of our ob- 
tions. The 
| the weight 
and not on 
e system we 
ontrol. Dif- 
system, the 
to the geo- 
te system in 
and terrain 
system. 
cil we have 
tem. Each 
antages, but 
lvantages of 
simple way. 
re must be 
the scale of 
ords on two 
| paper. The 
hem. It is a 
edetermined 
ts kind that 
d it is being 
ther method 
s. It would 
ANALYTICAL AERIAL TRIANGULATION, DISCUSSION 21 
be interesting now to hear why this has been 
abandoned and what the economy is. It would 
be of interest to see a publication in which the 
experience gained with this method is discussed. 
The second thing is that I would like to say a 
few words about the Herget method. The appli- 
cation of this has some disadvantages concerning 
the triangulation procedure and the condition 
equation, this extra factor, and improper linear- 
ising and weighting. But in spite of the objec- 
tions from the point of view of theory, one must 
not think that they are so serious that in practice 
when one’s photographs are good and one’s 
observations are good it will have much effect 
upon the results. You can still get good results if 
you have good photographs and good measure- 
ments. 
Discussion 
Prof P. WISER: Je remercie Monsieur Schut 
pour son intervention. Je voudrais également le 
remercier d’avoir respecté le timing, et je vais 
donner la parole à la première des personnes 
dont la participation au débat est prévue. 
Mr F. J. DoyLE: Mr Schut has pointed out 
in his paper some of the bases for a sound theory 
of analytical triangulation. One of the systems 
which he mentioned in his paper but not in his 
talk just now is that of Dr H. Schmid. One of 
the advantages to be gained by the use of high- 
speed electronic computation is the ability to 
invert a very large matrix. This presents us with 
the opportunity of solving a large number of 
photographs simultaneously. 
There are a number of advantages to this 
  
procedure and I do not wish to take time to 
point them out now. Dr Schmid is here and he 
has completed an investigation of a simultaneous 
solution of a large number of photographs. I 
have asked him if he would take the remainder 
of my time to show some of the results which he 
has obtained with this approach. 
Dr H. SCHMID: Mr Schut in his paper comes 
to one conclusion, namely that all the methods 
which he compares are based on the same 
geometrical condition, the condition that two 
rays intersect. In trying to derive a most 
economical analytical solution I found that this 
condition is somewhat cumbersome and leads 
to an undesirable observation equation, if the 
method is used to simultaneously evaluate a 
   
  
    
   
  
  
40 | m Ku 
mi A, ' 
m. Ku a 
i =mean square error of unit weight in microns Se A 
f 
3o 
2.0] 
s /, 
/ A 5 
Lo_ ^ Zz m, [m]s Ky-u -M-/0 
ug. C ç 4 
m, [m]: f, -u -M-/0 
m, [m]- K, u-M-07* 
uU smean square errorof uni! weighl in microns 
M: (scalefactor) 
0 5 10 i5 20 25 30 3 
No. of photographs 
Fig. 1 
ps 
in the strip 
40 45 50 55