ng 
se 
oo O 
DO 
iw 
  
22 
2-1 
2-0 
15. 
The effective area over which the measurement is made is much smal- 
ler than the area that appears to an eye to be covered by the scanning pattern. An 
MEAN SQUARE ERROR 
OF ELEVATIONS 
IN METERS 
  
i tutem 
be » Radius of circle 
02 of 04 .. 06... O0. 8mm 
  
r8 
I7 
  
L 
  
Fig. 7. 
attempt was made to deter- 
mine the size of the effec- 
tive area of measurement 
by comparing average ele- 
vations around each point 
with the results of the auto- 
matic stereo determination 
(Fig..7). 
From these re- 
sults it may be concluded 
that the diameter of the 
effective area over which 
the automatic stereoplot- 
ting device makes its 
measurement is about 0.1-0.2 mm in the scale of the photograph. It was found 
that the accuracy of the readings on the automatic stereo can be considerably 
improved in many cases by correcting the readings to an average height over the 
effective area. However, in a number of cases, an unexpected impairment of re- 
sults took place and so the mean square value of residual errors shown in Fig. 1 
is only slightly reduced by the corrections. The reason for this deterioration of 
some results when corrections for the surroundings were added has not been defi- 
nitely established. 
The above-mentioned deterioration of some results may be caused 
partly by another interesting phenomenon that affects the readings of the auto- 
matic stereoplotting device. It seems that the prevailing direction of prominent 
details has an effect on the "interpretation" the device provides for the elevation 
of points. An example of this is the case of two bridges that are used to control 
points in the International Test. One bridge is almost perpendicular to the photo- 
graphic base and its reading is almost correct and refers to the top surface of the 
bridge. The other bridge is nearly parallel to the base and its reading refers to 
the water surface. 
Automatic Contouring 
  
to the position of the contour line. 
The automatic stereoplotting device, once locked onto a given contour 
height, must correlate corresponding points in both photographs very rapidly and 
must move the tracing table accordingly. In human operation the drawing of 
contour lines is based not only on stereoperception but also on intuition. The 
operator, starting from a given point, feels where points of equal height are in 
the stereomodel and moves the measuring mark in that direction. As he moves 
the mark the operator checks its position continuously and when he notices that 
the floating mark is either too high or too low, applies corresponding corrections 
Proper perception and the interpretation of 
topographical forms is one of the conditions of continuous contouring by a human 
operator, because only then is he able to make a continuous decision on where 
next to move the plotting mark. This is the reason why the operator encounters 
difficulties in flat areas where it is impossible to anticipate the course of the 
contour line.