ints 
ment 
and 
uced 
0.1 
9e. 3 
ints 
are 
One 
es- 
hie- 
re- 
ixed 
tegy 
ap- 
loss 
heir 
ddi- 
the 
Af- 
work 
re- 
ted 
SEN- 
for 
ba- 
efer 
ital 
asu- 
data 
   
4 dd 
pos.1 pos. 3 pos.2 
0.15m 
plan view of camera positions 
  
figure 5: configuration of cameras: left 
plan-, right side view 
  
    
100 
50 
microns 
  
  
1 1 L L i I L L i 1 
5 6 
3 4 
control points 
  
stone 2 
lg] stone 1 
figure 6: standard deviations of control 
points in X/Y 
  
8 
microns _ 
  
  
  
control points 
  
stone 4 
| | stone 3 
|| stone 2 
  
li stone 1 
figure 7: standard deviations of control 
points in Z 
  
(camera slighly tilted) 
pos 3 
zi] pos 12 
dp 
0.6 m 
side view of camera positions 
o.1m 
  
  
  
  
  
  
  
stone 1| stone 2| stone 3 stone 4 stone 5| stone 6 
1l 7% 55 -63 96 65 -74 
o [2] -82 47 -79 -67 84 
9 8l M 45 81 -82 24 -54 
$4 72 6 -50 -81 -19 -27 
3 5| 26 40 4 54 77 58 
6| 84 -56 11 34 96 36 
  
  
  
  
  
  
unit = microns 
table 3: residuals of the introduced di- 
stances 
  
  
| sXY (micron) sZ (micron) 
Model 1 27 38 
Model 2 42 32 
Model 3 21 21 
  
  
  
table 4: residuals of 3 oriented models 
In /Reinhardt 1991/ the most impor- 
tant requirements to a DEM are performed 
(s. chapter 5). They concern 
the average accuracy of any point of a 
DEM, 
the accuracy of slope and curvature, 
the detection of gross errors, 
the completeness of morphological fea- 
tures and 
the fidelity and reproduction of cha- 
racteristic topographic shapes. 
The geometrical accuracy can be de- 
scribed by the following criteria as 
shown in /Balce 1987/ (error propaga- 
tion): accuracy of the control points 
(sap) after block adjustment, the resi- 
ST. after having fitted a model (sSyop) 
and the expected accuracy (correlated 
e.g. with the flying heigth in aerial 
photogrammetry (Smras)) - 
2 = 2 2 2 
S^cRIDPOINT S“AT+S“MOD+S MEAS (1)