6
PHOTOGRAMMETRIC ENGINEERING
a system whose sole aim is to facilitate their
recovery, and they are then measured in a
simple monocomparator, preferably one
which is provided with a reticle which permits
pointing by framing. It is certain that this
procedure is particularly economical; it is also
quite fast since point selection and point
measuring are not separated. It does not
eliminate the need to choose and mark the
images that will be used, but this preparation
is common to all the methods; at the most,
one can say that the points should be par
ticularly minute.
According to the articles published by the
D.O.S., the results of preliminary trials are
very satisfactory: if they are verified, they
will have made a serious and almost sacrile
gious attack on the fundamental principle of
stereophotogrammetry which says that only
by stereoscopic fusion is it possible to assure
the correct identification of corresponding
points that are not targeted on the ground,
with a precision that is considered axiomati-
cally to be higher than that of monocular
pointing. There is room for apprehension,
however, that the points whose coordinates
will be measured in the vicinity of the center
of more or less ill-defined spots, and which can
appear very different on different photo
graphs, may not be exactly homologous; but
it is possible that, because of the small dimen
sions of the spots, the error thus committed
will be on the order of a few microns and can
be considered negligible.
Great hope has been placed on automation
of procedures for selecting the points, and for
stereoscopic pointing, by use of diapositive
image correlators. But these image correlators
make only a relatively gross simulation of
stereoscopic fusion, and one cannot hope for
very much saving of time, since the coarse
matching would, without doubt, have to be
done manually. Besides the fact that the
equipment used in this work would be very
costly, the use of this equipment would not
eliminate the tedious work of preparation,
with the many manipulations of the plates
that would be required.
An entirely automatic solution, sometime
in the future, can nevertheless be envisioned;
it would consist of reducing all the photo
graphs to digital form, using a technique that
has already been tested: each photograph is
broken down into very small elements and
their photographic densities are recorded ac
cording to a scale composed of a number of
gradations, the analysis of density being
made, of course, automatically, by scanning,
with the aid of, e.g., an IBM analogue scan
ner. Starting with these data, recorded on
magnetic tapes, disks, or some other medium,
it would be theoretically possible to resolve
all the problems of aerotriangulation. Per
haps we will be set free from the idea of dis
creet points, the calculations being hence
forth concerned with finite areas, and ending
in the definition of complete topographic
surfaces, broken down into elements dx, dy,
by knowing the elevation and the photo
graphic density of each element, irom which
one could obtain orthophotomaps with con
tours directly. Thus, there would be realized,
at the same time as complete automation, the
synthesis of aerotriangulation and of plot
ting; but it is hardly necessary to state that
this will not take place tomorrow, nor before
the next Congress of the ISP.
References
Thompson, E. H., 1965, Analytical aerotriangula
tion: Bulletin de la Société Française de Photo-
grammétrie, no. 18, p. 3-8.
Inghilleri, G., and Galetto, R., 1967, Further
developments of the method of aerotriangula
tion by independent models: Photogrammetria,
v. 22, no. 1, p. 13-18.
Karara, H. M., 1967, Mono versus stereo analytical
photogrammetry; theoretical considerations and
experimental results: Photogrammetria, v. 22, no.
3, p. 99-113.
Eden, J. A., 1967, A new fast working approach to
analytical photogrammetry: The Photogram-
metric Record, v. 5, no. 30, p. 474-491.