86
b. Stereo video viewing with electronic shaping
of the video displays to remove geometric
distortions and scale differences between the
two photographs being observed.
c. Video monitor to display the area being
viewed on one photograph, the sum or differ-
ence of the video signals for two photo-
graphs, or an A-scope presentation useful in
accurate pointing of reseau and fiducial
marks.
. Image point coordinate data can be edited
before the coordinates are recorded. Editing
can be accomplished in many ways; recording
of some or all coordinates may be inhibited, for
example, under the following conditions:
a. The standard deviation of a coordinate from
the mean value is too large (replicated point-
ings can be accumulated and meaned within
the Comparator).
b. A two- or three-ray intersection in model
space fails by some amount larger than a
programmed threshold.
c. The point does not exist on one or more of
the photographs.
d. The pass-point identification number has not
been changed from the previously recorded
point.
. Supplementary data can be recorded together
with the image point number and the photo-
graph coordinates. This data may include the
standard deviation of each coordinate from the
mean for replicated pointings, the correlation
quality at final image match, any residual X or
Y parallax, the roughness of the terrain around
the pass-point, and the minimum scan area size
at final image match. This data is expected to
be useful in assigning a priori weights to pass-
points in analytical triangulation work.
. The Automatic Comparator increases pass-point
productivity in several ways:
a. No individual adjustments of image rotation,
magnification, or large translation are re-
quired for each pass-point.
b. No redundant pass-points must be observed,
since the editing routine automatically de-
tects and prevents recording of errors.
c. Automatic fine pointing by the correlator
speeds the conjugate point identification and
prevents operator fatigue.
d. Coordinates are recorded as photographic
coordinates, eliminating stage-to-photo trans-
formation during subsequent data processing.
e. Observed points can be assigned weights
based on image quality before subsequent
data processing takes place.
8. An additional feature of the Automatic Com-
parator is the extreme flexibility of system
operation. For example, the action taken by
the computer in response to each control-panel
switch is governed by a Fortran program; the
entire operation of the Comparator can be
changed by reading a different Fortran panel
routine and by relabeling the switches. For
example, with the addition of a coordinato-
graph, the comparator could be programmed to
function as a three-stage analytical plotter
somewhat similar to the two-stage Bendix-OMI
AS-11A or AS-11B Systems. With a standard
magnetic tape unit, the system could be pro-
grammed to automatically compile orthophoto
control tapes, digital terrain-model tapes, or
other special data forms based on stereo image
records. :
9. The Automatic Comparator should be of value
as a research tool in addition to its value in
production operations. For example, the ability
to use finite areas for pass-points with the elec-
tronic correlation equipment may prove to be
much more accurate than the use of discrete
points of image detail. Another potential area
for research is in the prediction of pass-point
quality from the pointing precision, correlation
quality, and terrain roughness about the point.
ORGANIZATION
Figure 1 shows the configuration of the Auto-
matic Comparator, and Figure 2 shows the functional
organization of the system. The central equipment is
an OMI TA3/P stereocomparator with several modifi-
cations, including: (1) addition of a flying-spot scan-
ner to each of the three photo stages, (2) inte-
gration of a high-magnification electronic viewing
system with the direct optical-train viewing system,
and (3) added operator controls and displays. The
electronic equipment associated with the modified
TA3/P consists of (1) a digital control computer
which performs the basic computation and automatic
control functions, (2) an interface which communi-
cates between the computer and other components,
(3) a scan generator which provides scanning signals
to the flying-spot scanners and electronic viewers, and
(4) correlation circuits which sense the errors in
matching pairs of conjugate images. A typewriter is
used for some computer input data and for output
recording. A card punch is used to record the final
point coordinates and other data on punched cards.
Special magnetic-tape units included with the com-
puter are used to store reseau and lens distortion
data.
FORREST, MOORE, AND SCARANO
"ue ed
el