The test results are summarized in table 3. 
MONO 
STEREO 
analyzed 
ground 
Pixel 
traffic 
buildings 
ground vege- 
waters 
topogr. 
terrain 
interpreta- 
examples 
pixel 
size 
networks 
tation 
forms 
model 
tion 
size 
roads 
paths 
large 
small 
de- 
forest 
other 
creek 
ditch 
gene- 
buil- 
tails 
ral 
dings 
rural 
orig. 
+ 
+ 
4 
+ 
4 
4 
+ 
+ 
0 
♦ 
44 
44 
4-4 
area 
1.8 m 
50 pm 
4 
4- 
4 
4 
4 
4 
4- 
4- 
0 
4 
44 
44 
44 
Vilie- 
3.5 m 
100 pm 
4 
4 
4- 
0 
- 
4 
4 
4 
- 
0 
44 
4 
4 
franche 
7 m 
200 pm 
4 
0 
0 
- 
- 
4 
4 
4 
- 
- 
4 
4 
- 
1:34000 
14 m 
400 pm 
28 m 
800 pm 
city of 
orig. 
4 
4 
4 
4 
4 
+ 
♦ 
4 
+ 
4-4 
+4- 
++ 
Hannover 
2.5 m 
50 pm 
♦ 
4 
4 
4 
4 
4- 
4 
4 
4 
4 
+ t 
44 
1:50 000 
5.0 m 
100 pm 
+ 
4- 
4 
0 
0 
4 
4 
0 
4 
4 
4-4 
+ 4- 
10.0 m 
200 pm 
4- 
0 
0 
- 
- 
4 
0 
- 
0 
4 
4 
0 
15.0 m 
300 pm 
4 
- 
- 
- 
- 
4 
0 
- 
- 
4 
0 
- 
20.0 m 
400 pm 
- 
- 
- 
- 
- 
- 
- 
- 
- 
- 
- 
40.0 m 
800 pm 
- 
- 
- 
- 
- 
- 
- 
- 
- 
- 
- 
- 
Island 
orig. 
4- 
4 
4- 
- 
4 
4 
0 
0 
4-4 
+ 4 
44 
of Sylt 
6.3 m 
50 pm 
4- 
- 
0 
- 
4 
4 
0 
0 
4- 
44 
44 
1:125000 
12.5 m 
100 pm 
4- 
0 
0 
- 
0 
4 
0 
0 
4 
4- 
4 
25.0 m 
200 pm 
0 
0 
- 
- 
0 
4 
- 
- 
0 
0 
- 
50.0 m 
400 pm 
- 
- 
- 
- 
- 
- 
- 
- 
- 
- 
- 
100.0 m 
L 
800 pm 
- 
■ 
- 
“ 
• 
- 
- 
- 
- 
- 
- 
Table 3: Interpretation results 
The following topographic features have been extracted: roads and paths, 
buildings, vegetation (forest, trees and shrubs), creeks and ditches 
and topographic forms. 
The ability to interpret these features in the different images has been 
indicated by the following symbols: 
++ excellent; + good; o satisfactory; - not satisfactory; 
blank not contained in image 
The interpretation of the original images is satisfactory for all cases. 
This is, of course, to be expected since small aerial photographs are 
generally used for mapping. There is one slight exception in that buil 
ding details such as ledges etc. cannot be detected in scales 1:125 000. 
This is, however, not significant for the required map content 1:50 000. 
In order to depict such building details pixel sizes between 2 and 5 m 
are at least required. 
The most significant result of the test is that stereo-observation is by 
far superior to monoscopic observation. The stereoscopic advantage in 
creases nearly proportional to the base height ratio. 
The comparison between the different areas photographed shows clearly that 
the interpretability of houses, roads, land forms, creeks and ditches 
also varies with the type of the terrain since urban areas, rural areas 
and uncultivated areas have generally different types of housing, roads 
etc. even though they are of the same topographic feature category. 
The most critical features are the buildings. To map them monoscopically 
a pixel size of at least 3 m is required. Under stereoobservation a pixel 
size of 6 m will suffice. This is the requirement for urban areas to map 
at 1:50 000. 
To map roads and vegetation this requirement may be slightly relaxed. 
The results of investigating digitized images at 25 pm intervals were 
identical to those in the original. Therefore they are not contained in 
the table. This just proves that the observer's eye cannot resolve better 
than about 5 to 10 lp/mm. 
Interpretation under stereoobservation is possible for 50 pm and 100 pm 
pixels. One may see stereo also for 200 pm and 400 pm pixels but this 
does not aid in interpretation any more and does not permit contouring. 
383