Se
ACCURACY OF AUTOMATIC CONTOURING
FROM 1:3000 PHOTOGRAPHS
-06m
- 0-5
A Differences: Automatic- A7
- 0:4 O ,
- 03 S
m 0:2 o.
Joan Theoretical accuracy
FO Sor e oor e m, =(0-1+0-18 tanoc)m
Slope in %
Average height error of contours
I 1 T
5 Io 15 20%
Oo
Fig. 10.
proposed a modification in the scanning arrangement to produce an improvement
in the contouring accuracy of rugged terrain. A further overall increase in
accuracy can be expected from the adaptation of the automatic plotting device to
the non-analglyphic plotters.
Automatic Contouring from 1:50, 000 Photographs
As for the large-scale photographs, automatic contouring from a small-
scale stereogram was compared with the plot produced on a Wild A-7 Autograph.
Both the automatic contouring and the A-7 plot were carried out at a scale of
1:10,000. The contour interval was 10 m. Identical points were used for the
absolute orientation in both these experimental plots. The distinct systematic
difference between the plots makes a precise evaluation of the accuracy of the
automatic contouring rather difficult. In addition, there are occasional forests and
built-up areas that add to the difficulties in accuracy evaluation. It was decided
however to present the differences between the A-7 and the automatic plot
uncorrected as they stand.
Figure ll shows the original automatic plot reduced to a scale of
1:25,000. For evaluation purposes the original plot was corrected by the drafts-
man who drew in contour lines to follow the local terrain forms where the pattern
of automatic contour lines were confused.
Because of the relatively small scale the corrected automatic plot
(Fig. 12) and the A-7 plot reduced to 1:25,000 (Fig. 13) are presented separately.
Comparison between A-7 and the automatic contouring was carried out
on 1:10,000 plots by using the same technique as for large scale contouring.
Because of the constant variation in slope only one contour line (140 m) was used
in its entire length. For the evaluation of other contour lines only those portions
were used that traverse open areas of more or less constant slope. The results
are listed in Table III.