650 - 
Some experiments on the utility of the method in aerial photograph interpretation w 
followed suit, however, (FONTANEL et. al. , 1966; CHEVALLIER et. al. , 1968, 
GUY, 1969; NYBERG et. al. , 1971). Since the techniques seem to be promising as 
th.3 
for those fields of geomorphological image analysis related to the directional 
properties of landforms, a brief description of the methodology and of the appli 
cation is given in section 4 of this paper, using some examples related to the An 
analysis of dune patterns and crevasse patterns of glaciers respectively. of i 
tl tc 
The examples of quantitative image analysis given in the following sections all 
relate to the spatial distribution of landforms using relief and density character- 
eli: 
istics and directional properties as criteria . 
eas 
2. RELIEF ANALYSIS aS 
sup 
So far, topographic maps have generally been used for the purpose (HORMANN, is ] 
1968, 1969; STRAHLER, 1968). Relief analysis of aerospace imagery is of par- drc 
ticular interest if stereoscopy can be obtained. Qualitative analysis can then be the 
supported by measurements of height differences, slopes, ruggedness, and nu- slo; 
merous other means of quantitative relief analysis. If a systematic relief ana- 
lysis and automation is aimed at, digital terrain models may be an answer. An- nu 
other approach in such cases, however, is through the construction of a number se j ( 
of parallel and equidistant profiles. The dropped line plots produced in conjunction jic 
with orthophotography are particularly valuable in this respect. Some elaboration d 
is thus justified in the context of this paper. e j e] 
The output of orthophoto production is a differentially rectified reproduction of GUf 
the aerial negative giving correct geometric relationships of the image (see also val] 
VISSER & v. ZUYLEN, 1968). The dropped line plot produced in conjunction gives 
the vertical (Z) dimensions of the model. The contourlines can be drawn by hand 
in \ 
on the basis of the dropped line information or may be obtained by the automatic 
K i$ 
production of contour line tracings. Stereoscopy with orthophotos is normally 
T6C 
achieved by fusing an orthophoto with a normal photo forming a pair. Since the 
to p 
orthophoto is enlarged, either an enlarged normal photo should be used for the 
purpose or the scale difference between orthophoto and normal photo should be ^nc 
compensated by differential magnification. Since in this case the X-parallax am I 
producing the 3-dimensional effect is not the additive effect on two photos but re- c °d ( 
suits from one (the normal) photo only, the stereo effect will be reduced, and tain 
will even be zero in the principal point of the normal photo. Important deforma- eve 
tions of the model, affecting the relief interpretability, thus result. Another of r 
resi 
is t]