24 
From the general law of error propagation it is easy to determine the 
standard error of the corrections of the coordinates in the selected points. 
3h 
Using the relations b’S = b = d= — 
= 
© 
transformations and with minor approximation the standard errors 
we find after some simple 
m, = m, = 0 
1 va 
ue. DM ——2 94 — ei —— 15$ 2 AC 
My, = My, = M, = My = M, =M,., =3&% (69) 
2 T3 
  
/ 
m, =m, =m, =m, =S§, | 11 
In these expressions s, is the standard error of unit weight of the 
y-parallax measurements. Evidently, the weight distribution after the 
correction with d», and dx, can easily be taken into account when the 
error propagation within single models or within the strip is being 
investigated. Similarly the corrections dp and dw can be treated. 
1.2. Basic formulae for the triangulation 
1.21. The principles of the triangulation procedure 
The strip to be triangulated is assumed to consist of the individual 
photographs — 1, 0,1, 2... .. n — 1, n,n + 1or of n + 2 photographs 
fig. 1). Only two control points are assumed to be present. one in the 
e v 
photographs —1,0 and the other one in the photographs n, n + 1. The 
control points are further assumed to be suitably located with respect 
to the accuracy of the intersection procedure, see diagrams 1 and 2. 
In fig. 1 the control points are denoted with small triangles. 
Each pair of photographs is treated separately in accordance with 
the procedure as demonstrated by HALLERT 1957 a and 1957 b. In 
each pair of photographs the image coordinates of the available control 
points and of at least four transfer points (3, 4, 5 and 6), see fig. 1 are 
measured. The transfer points 4 and 6 in each pair must be identical 
with the points 3 and 5 respectively in the pair to the right. It is of 
course very important that the transfer points are chosen with great 
care. If the image coordinates of each photograph are to be measured 
individually (in a single image comparator) the transfer points must be 
chosen in image details which can be identified with high precision in 
the adjacent photographs. 
If the image coordinates are to be measured simultaneously in two 
overlapping photographs (in a stereocomparator) artificial transfer points 
according to EKELUND 1950 and 1951— 52 can be used, see fig. 5.