This
icci-
atic
k of
ech-
e at
sub-
that
nts,
e of
| by
lard
)no-
ints
by
the
rec-
ase
rom
fer-
t it
xes,
ons.
| be
for
phs,
ıted
any
een
vel-
by
ter-
plet
lets
gu-
"ec-
of
ery
ent
int
or
on-
let
RADIAL TRIANGULATION, ROELOFS 73
methods, the accuracy of which is indeed of the same ma
; gnitude as R.T. with vertical
photography. The methods and conclusions are explicitly claimed to be valid also for
numerical R.T.
It is the present author’s opinion that the stereotemplet system does not fall in the
category of R.T., a stereotemplet representing rather a group of points on an arbitrary
scale than a set of directions as is essential for R.T. In this report this interesting stereo-
templet system, which was suggested by Scher, [12], and reached its highest perfection
in the LT.C.-Jerie block adjustment method [13], is therefore only mentioned because a
combination with radial templets may be a practical suggestion. When, by lack of vertical
control and sufficient hydrographic features, the level solution of one or more models is
doubtful, it is normal practice to exclude these models from the templet assembly. It
would be worthwile to investigate the feasibility of filling the gaps thus produced and
strengthening the assembly by using radial templets for the models in question.
4. General Remarks.
The fact that the mechanical R.T. has so successfully superseded the graphical R.T.
is not only due to an easier handling of a single strip, but principally because it achieved
what graphical R.T. was not able to: adjustment of a block of strips. It is in fact a
block adjustment method, which requires simpler equipment than any other system. It
is to be expected therefore that it will maintain its position, although it suffers from the
competition with the stereotemplet system which is more universal, but requires more
expensive equipment.
There are now available various methods of block-adjustment — of which the LT.C.-
Jerie system [13] is undoubtedly the most attractive - which ensure a much higher pre-
cision. This means that the high precision numerical R.T., if combined with block adjust-
ment, gets a new chance.
In most cases principal point R.T. will give sufficient accuracy, on the condition that
the photo-flights are made with modern auto-piloted aircraft, with a crew experienced
in photo-flying and especially instructed about the importance of small tilts. Actually it
is only exceptional to-day that air photography, made with care, shows tilt values of
more than 1° only.
If still greater accuracy is pursued, nadir point R.T. comes into the picture. It
requires the determination of the camera inclination for locating the nadir point on the
photograph and for computing corrections to the directions measured. There is a multi-
tude of methods to deduce the camera inclination from the measurement of coordinates
or parallaxes in the photographs. These methods require only simple equipment, obser-
vations of moderate accuracy and elementary computations, which makes them suitable
for a production-line system. They are however oniy a make-shift, destined to be with-
drawn as soon as a direct method of tilt determination is applied.
While the present-day pendulum-gyro-systems already reach an accuracy of 10'-15',
which for most purposes of R.T. is excellent, it is to be expected that the modern extreme
precision inertial stabilizing instruments will become available one day for photogram-
metric purposes [14]. These methods, giving an accuracy of 3’ will be of great usefulness
for photogrammetry in general and for radial triangulation in particular.
Literature.
[1] Roelofs, R., The Future of Radial Triangulation. Photogrammetria 1955-56, p.
[2] Dmochowski, S., The Mean Square Error in the Determination of Plane Co-
ordinates of Some Points of a Radial-Triangulation Chain. Proceedings of the
Institute of Geodesy and Cartography. Warsaw, 1957. (Polish, English sum-
mary).