means 
trans- 
1d co- 
justed 
tment 
order. 
ber- 
ndard 
errors 
sal he 
lon of 
chief 
1, etc. 
, mea- 
stage. 
ncipal 
eight 
) that 
. The 
ft be- 
plane 
rator. 
uring 
ments 
cecom- 
der is 
mea- 
leasu- 
llaxes 
val of 
| mm 
,efore 
mea- 
  
surements. The principal point has been computed as the average of the two opposite 
fiducial marks. 
For relative orientation three points have been selected close to the Western edge 
of each photograph. These points have been measured in each model where they oc- 
curred. This gives nine points for relative orientation of each model. In many cases 
where the longitudinal overlap is considerably more than 70%, three more points 
have been chosen along the Eastern edge of a model. In a number of models, where 
the overlap was not quite so large, and the nine points covered the whole width of the 
model, three additional points have been measured farther inside the Eastern edge. 
As a result twelve points have been used for relative orientation of most models of 
the six Northerly strips, and nine for relative orientation of most models of the two 
Southerly strips. 
In addition one tie point with the strip to the North has been measured in each 
model. As tie points, points have been chosen that were used for relative orientation 
of the Northerly strip. The ground control points have been measured also, 1n most 
cases in one model only. Each point was measured three times stereoscopically, and 
the average of the readings was recorded. 
2. — Strip triangulation 
The strip triangulation has been performed on an IBM-650 electronic computer, 
using the N.R.C. method of analytical aerial triangulation. This method consists in 
orientation of each photograph of a strip in succession with respect to the preceding 
one. In the orientation procedure relative orientation of the photograph and scaling 
of the resulting model are performed separately. 
The program for strip triangulation converts the parallaxes measured for the right 
photograph to coordinates and applies radial corrections for lens distortion, earth cur- 
vature and refraction. The strip triangulation is performed with respect to an arbitrary 
rectangular coordinate system, choosing arbitrary coordinates for the projection centre 
of the first photograph and choosing the unit matrix as the orthogonal matrix for 
orienting this photograph. 
Each strip has been triangulated from West to East. The first model in each strip 
has a base component b, in the strip direction of 4000 m, following models are scaled 
by means of the three points along the Western edge. 
With the present program, generally strips have to be triangulated twice. The 
first triangulation serves to identify reading errors. After identification of reading 
errors and correction of the errors or removal of the affected points the triangulation 
Is repeated. 
9 
3. — Strip transformation and levelling 
Since great height differences occur in the area covered by the block, the strips 
have been levelled roughly before starting the block adjustment. As a result of the 
earth curvature corrections, the triangulated strips have only small longitudinal bends. 
This is shown by the Z-coordinates of the projection centres. Therefore, for levelling 
of the strips a three-dimensional linear transformation, without deformation, has been 
employed. 
Strips 1 and 8 contain sufficient known ground control points for this transfor- 
mation. The remaining strips 2 to 7 have been levelled in strip direction using the 
heights of the projection centres and estimated heights of a river in the Eastern part 
of the block. This river flows past known ground control points in the N-E and S-E 
corner. An approximately parabolic rate of change of its fall has been assumed.