16. 
In automatic contouring the process is similar, although different prin- 
ciples are involved. The instrument obtains signals that indicate the location of 
points of selected horizontal parallax, and from these signals the tracing table 
draws contour lines. Inertia, which may reach appreciable values at the high speed 
of the tracing table, has the same effect on the contouring as the wrong judgment 
of the operator for the anticipated course of the contour line. The effect of inertia 
is controlled continuously by new signals that direct the tracing table along the 
proper contour line. Since the present instrument cannot interpret the content of 
photographs, automatic contouring is affected by vegetation and artificial objects 
on the ground. However, there is an averaging process built into the electronic 
circuitry of the automatic plotting device, and so the effect of scattered trees and 
other small objects is quite negligible. 
On the other hand, since on large-scale photographs isolated groves of 
trees or other objects will cover relatively large areas, the device may start to 
contour those features and the help of a human operator may be needed to over- 
come this difficulty. 
Automatic Contouring from 1:3000 Photographs 
Figure 8 is a photographic reduction (1:3. 3) of an original plot produced 
by the automatic plotting device at a scale of 1:600. The scale of photographs was 
1:3000 over test area A, which consists of relatively flat cultivated country with 
occasional farmbuildings and single trees. The speedof automatic contouring was 
twice the peak speed of contouring by a skilful human operator. The contour 
interval is 4 mm in the model scale, or 2.4 m on the ground. 
A detailed analysis of the automatic contouring reveals that, with the 
exception of the very flat areas, trees and houses cause some confusion in con- 
touring. Actually, only some are marked on the plot but the missing trees may 
easily be located on the plot by comparison with the aerial photograph reproduced 
in Fig. 1. The only unexpected and unexplainable contouring error occurs at the 
spot marked "interference of houses’ by the operator. At this spot the instrument 
first became confused because of the houses and then continued contouring for a 
while in an apparently normal manner that was actually entirely erroneous. 
Otherwise the contouring is surprisinglyaccurate and depicts even very 
minor terrain forms. Figure 9 shows a comparison of automatically-plotted con- 
tours with contours plotted on a Wild Autograph A-7. 
This illustration must be interpreted carefully because some of the dis- 
crepancies between the two plots are obviously of a systematic nature and may be 
the result of differences between the types of plotters rather than between auto- 
matic and human contouring. The A-7 contours are obviously smoother, but we 
would hesitate to make a statement as to which contour is more accurate. When it 
is assumed, for instance, that the 188-m A-7 contour represents the true contour 
line, the maximum deviation of the automatic contour amounts to about 0. 6 m on 
the ground. If the spots where the device loops, etc., are disregarded, the maxi- 
mum errors within the plot reach no greater value than 1.2 m. In order to offer 
more precise figures the deviations of all contour lines from A-7 contours were 
determinedby a simple procedure that has been used on various occasions at NRC. 
The area between the reference contour line and the contour line to be compared