118 CONVERGENT VERSUS VERTICAL PHOTOGRAPHY, SCHERMERHORN
We must conclude here to an unfavourable precision of the convergent camera com-
pared with the RC7. To obtain the same precision we should reduce the scale of photo-
graphy considerably, which reduces then the economy of the convergent system. The
economy suffers also from the long time per model for relative and absolute orientation
(Is this due to the participation of institutes with operators without experience with
convergent-photographs?)
Whether however these results are significant, notwithstanding the large number of
participants is not sure. Comparing the 17 u results of strip I with the 12 u obtained with
the RCT photography Oberriet (Exp. S.LP. and O.E.E.P.E.) we can accept this differ-
ence as non-significant. The difference however between strip I and III is much worse.
It is very likely that the 344 are largely due to the influence of errors in the inner
orientation for which convergent photography is very sensible, in particular in moun-
tainous terrain.
Then one conclusion could be that not sufficient precautions were taken in this re-
spect. The other conclusion that this sensitivity is a disadvantage of the convergent system.
There is some ground for this conclusion if we look at a second, recent experiment.
It is carried out by the German Institut für Angewandte Geodäsie in Frankfurt. Al-
though the results are not yet final, they are of sufficient importance to be mentioned,
in particular to put the Monte di Revoira tests in the proper light.
The following photographs with their restitution have been used:
| Scale | Scale | Number of
lvi ; . ;} e | € | .
Flying height | photogr. model | models
| |
a. 400 m 1:1,900 | 1:1,000 | 8
b. 800 | 1 : 3,900 | 1 : 2,000 | 3
c. 1200 1 : 6,000 | 1 : 3,000 2
The photographs have been taken vertically and convergently with the Zeiss 2 X
RMK 21/18 convergency 308.
The orientation was carried out partly optical-mechanically, partly numerically.
The total orientation time was abt. 9 hours per model which must be considered to be
entirely abnormal. The coordinates have been obtained in two cycles. The area contains
about 900 check points.
The following preliminary results have been obtained:
I. Convergent photography: mean square errors in scale of the model: m,-—14y, m,-—1" y.
The mean square errors in the plane of the negative are: m, 5 Tu, M = Ip.
II. Vertical photography: mean square errors in the scale of the model:
m, = 13 u, m, — 26 y.
The mean square error in the plane of the negative: m, = Tu, M, 13 u.
The conclusion is that the mean square errors in the planimetric coordinates
Mm, =m, = my, are equal for vertical and convergent photography. The mean square
errors in height are in accordance with the proportion between the base height ratio in
the two cases, as could be expected within this range of base height ratio.
Although in the Netherlands the convergent camera came recently into use again,
we cannot provide results of precision tests which can be considered as final and reliable.
The execution of aerial triangulation of convergent photography has resulted in sys-
tematic distortions of each of the convergent models. Since there was, originally, a
suspicion that this could be due to the lack of adjustment of the Stereoplanigraph C8,
part of these triangulations were carried out in the Wild ATK of KLM Aerocarto. The
hav
wit]
of
the
wel
wh
of
wh
wil
ha
thi
lec
WI
pu
bu
be
pe
th
pa
m
re
m:
