5
ACQUISITION OF DATA IN AEROTRIANGULATION
In the absence of pretargeting, one can use
natural images, such as the corner of a build
ing, the centerlines of crossroads, etc., but
such details are rarely distinct and sharp, and
they may have different appearances on
different photos, making the measurements
uncertain. Recourse to stereoscopic vision,
and consequently to a stereocomparator, is
quite indispensable, but this alone is not
enough to assure the transfer of a control
point from one strip to another.
An attractive-appearing solution is the
marking of the points on the plates, using an
instrument like the Wild PUG or the Zeiss
Snap Marker. There are two ways in which
they can be used:
1. Each point can be marked on only one
plate of each strip, the pointing being done
stereoscopically in a stereocomparator. But it is
not very practical to make a precise stereo
scopic pointing if one of the points is already
marked, unless, as pointed out by Prof. E. H.
Thompson, the shape and dimensions of the
stereoscopic index are exactly adapted to those
of the mark; in fact, it is well known that point
ing by superposition of identical images is far
less precise than pointing by bisection or by
framing.
2. The second manner of operating consists in
marking all the points on all the plates. To carry
out this operation, it is necessary to associate
one plate, on which the point has been chosen,
successively with all the other plates upon which
the point appears, and to make a very precise
stereoscopic pointing on each, the first plate
being marked last. This amounts to the same
thing as in the case of pretargeted points, and
the measurements can be made very rapidly on
the stereocomparator. Prof. Thompson has
criticized this method on accuracy and economic
grounds. It is certain that the fact that there are
in reality two pointings, one stereoscopic and
the other monocular, reduces the precision to
some extent, to say nothing of the errors intro
duced in marking the points or of the greater or
lesser suitability of the mark made in the emul
sion. On the other hand, regarding the economic
considerations, one can hardly dispute the fact
that the marker-monocomparator combination
is appreciably less costly than most of the mono
comparators now on the market, especially since
a considerable number of markers can be associ
ated with a single monocomparator. Prof.
Thompson has justly stated that the slowest and
most demanding operation is the association of
the plates, two at a time, and their arrangement
for stereoscopic examination: once this has been
done, it is not much more difficult to measure
the coordinates on the stereocomparator than to
do the marking. But the difficulty is that if one
adopts this method, the measurements must be
made point by point rather than plate by plate;
consequently, each plate must be put in place
for measuring as many times as the number of
points it contains, and, each time, it is necessary
to point and to measure the coordinates of at
least two fiducial marks, so that the measure
ments can be related to the same system of axes.
One can perhaps avoid these repeated measure
ments if the positioning of each plate on the
plateholder can in some way be recaptured with
a precision on the order of a micron. But one
cannot escape the manipulations necessary to
associate the plates two at a time, which is a
source of time loss incompatible with the
efficient utilization of so expensive an instru
ment.
The method of photographic recording ap
pears capable of solving the irritating problem
of how to transfer points in aerotriangulation.
It is on the Nistri ТАЗ stereocomparator
that one finds the principal application of
this technique; this instrument allows one
to photograph on 35-millimeter film the
image seen through the oculars—that is, the
point with the reticle superimposed. This idea
has been adopted in the SOM stereocom
parator, with the difference that a reversed
positive made from the film can be introduced
in the instrument, examined in one or the
other of the oculars, and fused stereoscopi
cally with the corresponding area on the plates
being measured. This resolved, at least
theoretically, the problem of transferring
from one strip to another. But, as a matter
of fact, this possibility has never been utilized
in practice at the IGN because of the com
plications introduced in the operational pro
cedure, the insufficient capacity of the 35-
millimeter film magazine, the difficulty of
obtaining a correct exposure, and the addi
tional burden imposed on the photographic
laboratories. The differences in scale between
strips, entailing inequalities in the apparent
diameters of the photographic mark and the
instrumental mark, would probably have re
vealed difficulties that appear more serious, al
though not insurmountable. However, aban
doned by the IGN, the idea was momentarily
picked up by the U.S. Geological Survey. Ac
cording to information submitted by Marvin
B. Scher, consideration was given to applying
the idea to an instrument of the monocom
parator type; but this project, unfortunately,
does not seem to have been brought to frui
tion.
A novel and somewhat surprising solution
has been experimented with in the last few
years by the Directorate of Overseas Surveys;
it consists of using as control points natural
details in the photographic image, the novelty
lying in the fact that these details have very
small dimensions (a few tens of microns).
They are selected according to a somewhat
complex procedure in which, after stereo
scopic examination under moderate enlarge
ment, the final choice is made under very
great enlargement, but without precise stereo
scopic fusion. The points are then marked by