n re- 
ging” 
uch more 
; of topo- 
automate. 
lanimetry 
accurate 
relieving 
king con- 
races out 
ssistance 
0 percent 
» the pro- 
corrected 
errain re- 
exposure. 
s a sub- 
purposes, 
'antage of 
available 
ons where 
wn plani- 
y be used 
f features 
afting and 
o-plotting 
ice ortho- 
rinting C. 
ion would 
of the 
inting 
it out- 
'inting 
gener- 
ws the 
SR. T. 
ain is 
ing C. 
s this 
{xpos- 
model, 
ed for 
probably 
ecause of 
omat on a 
have pro- 
a Ouro 
available 
  
vision for relative orientation by movement of 
either projector alone. This would require motor 
drives to all six projector axes and for Bx, By 
and Bz motions also. Motors for common orient- 
ation of both projectors may also be required. 
The operator would normally select the axes or 
motions used, although an automatic program 
could probably be devised, : 
Stereomat can also assist triangulation 
by the semi-automatic locating and recording of 
significant points in the photographs. For the 
automatic location of control points, a small 
supplementary photograph of the terrain is re- 
quired with the control point at the centre. The 
supplementary photograph could be cut from a 
paper print, or preferably, printed from the 
negative on to a small glass plate. 
An automatic triangulation instrument would 
be provided with a separate scanning means for 
supplementary photographs. To set up control 
points the procedure would be as follows: 
1. Relative orientation would be completed. 
2. The photo-cell of the separate scanner would 
be connected to the correlation unit in place 
of the photo-cell for either the left or right 
plate. 
3. The floating mark or pattern centre would be 
set to the known coordinates of the control 
point. 
4. The X and Y parallax signals developed by 
the correlation circuits would be used to 
actuate a common orientation axis of both 
projectors, or to adjust Bx, as selected by 
the operator. 
5. The procedure would be repeated for other 
control points, the parallaxes being auto- 
matically cleared in each case by an ap- 
propriate common orientation or base adjust- 
ment. 
An alternative procedure would be to ad- 
just only one plate to the control point, thus 
avoiding the requirement for common orientation 
motors. Normal scanning would then be resumed, 
and the second plate again brought into relative 
orientation. 
Automatic recording of pass points is also 
possible. Pass points selected by the operator 
could be printed by a small C.R.T. similar to 
that used for producing ortho-photography. The 
resulting print can then be used to set up the 
plotting instrument in a manner similar to that 
just described for control points. 
PERFORMANCE FACTORS 
AUTOMATIC STEREO 
be expressedas the fraction of the average model 
area that the system will fail to plot automati- 
cally. 
Accuracy in turn can be divided into static 
accuracy and dynamic accuracy. Static accuracy 
refers to the system at rest, and is primarily a 
function of the parallax sensing operation. Dy- 
namic accuracy refers to the additional errors 
that are introduced as the result of the motion 
during profile or contour tracing. 
Plotting speed and accuracy are inter- 
dependent with Stereomat, and increased speed 
can always be obtained at the expense of re- 
cord smoothness andaccuracy. For a given accu- 
racy, however, the plotting speed is limited 
fundamentally by the rate at which information 
is extracted from the stereo photographs. Scan- 
ning the images at higher speed increases the 
information rate, but also raises the frequency 
of the image signals, thus increasing the diffi- 
culty of separating them from the spurious 
"quantum noise" signal also present in the out- 
put from the photo-cells. 
Quantum noise refers to irregularities in 
the image signal owing to the corpuscular or 
quantized nature of light. At very low-light levels 
the rate at which light quanta or photons reach 
the photo-cells may be low enough for them to 
register individually. The resulting roughness 
in the output signals tends to obscure the image 
signals, and to introduce irregularities in the 
data signals from the correlation circuits. These 
irregularities are transferred to the profile or 
contour records, and produce an erratic Jiggling 
or oscillating of the servos. 
Increasing the amount of light reaching 
the photo-cells reduces quantum noise, and there- 
fore allows a higher scanning velocity to be 
usefully employed. There are physical limits to 
the brilliance of the scanning spot however, and 
Stereomat approaches these limits closely. 
Optical aperture is, therefore, a factor governing 
CIRCULAR SPIRAL. © 
RADIAL. ROSETTE 
  
  
TV RANDOM cé 
Three performance criteria for an automatic 
plotting system are: speed, accuracy, and falli- 
bility. Speed and accuracy are generally ex- 
pressed in numerical terms, and fallibility could Figure 6. Scanning Patterns 
TT FTOLEHSDUI UXL LIL CU VX.LXQ Gk AAA whe NE TTS 
Canada. 
1ed, 
sst 
1at 
18 
usly 
f 
lates 
a of 
the 
of 
o 
atrol- 
ohoto- 
t at 
2e8Ssary 
ang 
achine 
spot 
as 
ns 
at 
for 
efore 
tions 
3 
will 
lace