; the
ation
ea is
»jek-
Ings-
urze
von
ction
ment
."Un
is de
urry
as it
coat-
re it
Vhite
| the
s we
the
side
lies
o its
t by
n of
etric
tific
kills.
jhink
w, it
egin-
AUTOMATION IN PHOTOGRAMMETRY, ESTEN 139
nings. Later speakers will tell you what is presently available and finally we shall tell
you what we think the future holds.
Early patents in the field of automatic contouring date back to the early 1930’s, but
no substantial experimentation was conducted until 1950 when the United States Army
Corps of Engineers contracted Bausch and Lomb Optical Company for the study of an
automatic topographic scanning device. Figure 1 shows the equipment developed by that
company by June 1951. It consisted of a small conventional multiplex unit supplemented
by a scanning chamber which was transported over the model area by X and Y screws.
The instrument operated in the profile mode. The scanning chamber contained two
photomultipliers, one to scan a small dimension in X on each diapositive. The electrical
signals produced were compared in phase and the resulting signal used to drive the
chamber in Z. Thus when the right wave led the left, the scanning unit was too low and
a signal was generated to drive the scanning chamber up onto the profile. When the right
wave lagged the left, the scanning unit was too high and the scanning chamber was
driven down onto the profile. The equipment proved that automatic contouring was pos-
sible.
The next instrument shown in Figure 2 was an adaptation of the previous instrument
by Pickard and Burns, Inc., for the Corps of Engineers. This instrument utilized
automatic gain control so that the signals from the photomultipliers might be compared
in amplitude. This instrument scanned the model at one elevation and in one pass
produced only the contours for that elevation. This instrument indicated the great speed
that might be expected with automation.
The next contractor, Hycon Manufacturing Company, was more ambitious and at-
tempted to contour directly by automatie means. This instrument is shown in Figure 3.
We realized that we required a stronger signal with which to work and therefore switched
to a Kelsh Plotter. Although this attempt was not particularly successful, Hycon did
introduce the idea of area correlation (one which finds almost universal use today).
Rather than attempting to match points on a line, area correlation matches points within
a small area. Search for detail to match tended to make the size of this area larger, while
the question of relief discrimination tended to reduce its size.
Lack of precision in the scanning process directed our efforts next to the devel-
opment of equipment to traverse the model volume and scan mechanically the stereo-
projected images. The instrument developed is shown in Figure 4. It features a tilting
platen to compensate for scale distortion due to slope. Development of this unit with a
new correlation unit is currently underway.
A semi-automatic map compilation system is also being constructed for use by the
Army Map Service. This development shown in Figure 5 will be presented in more detail
in the proceedings of Commission IV.
While this work with projection plotters has been underway, the Corps of Engineers
has also been conducting investigation on what is called an Automatic Map Compilation
System. These operations differ from the preceding by the fact that no optical model is
produced. The stereo-pairs are scanned only in plan and the necessary transformations
and interpolations made by high speed digital computer. Figure 6 represents the early
phase of this investigation. At the present time this investigation is being conducted both
along analog and purely analytical lines.
We shall not be late! I am not about to pull a rabbit out of a hat, but I do wish to
introduce to you now some of my colleagues who will tell you about other advances in the
automation of photogrammetry.