tion, unexpected at the present scale of photographs and by no means meeting the measuring accuracy.
Therefore errors in the control points or defects in the photographs (film) were assumed. To find the sources
of errors additional vertical control points were determined in the field pertaining to several pairs of
photographs. Thus every pair of photographs covered about 8—12 known spot elevations. The absolute
orientation, however, still revealed major irregular deviations in these points. The discrepancies in
bridging the borders of successive models therefore cannot be explained by incorrect elevations of control
points alone, for these errors ought to be systematic ones, otherwise we should have identified the wrong
control points.
In order to determine the magnitude of the individual errors we used four stereoscopic models. Every
two of them have one picture in common. We selected about 50 points in the area twice overlapped by
contiguous stereoscopic pairs. Then two operators of the Plan und Karte Company at Münster (Germany)
independently observed each stereoscopic model by means of the Stereo-Plotter (the scale of stereo-
models being 1 : 7,500). At first the given vertical control points were measured in two runs. Subsequently
the points located in the border zone were observed, disregarding the fact that one or more of these
points were vertical control points too. In the first measurements diapositives were used. In order to
guarentee that no additional errors occur in reprinting the original films, these films were later on measur-
ed by the first operator too. Consequently we have at our disposal three plottings each of the four stereo-
scopic models (i.e. a total of 12), comprising two sets each.
[o Ey) Os!
O 53
Q9:
055 Qi Q5;
© 2
07
O 56
© ^4
05? 05 © un
O58
QA: O6!
Oso
© 544 Q9; orn © 564
lo ry, Qe
O64
O 66 065
O69
© 45/9 99 O 68
© 1776
© 63/2
orn Q70
Qu: Q
Fig.2 — Distribution of control points and border points in the stereopairs 80/81 and 81/82
