of the land 
1suming 
y individual 
se boundaries 
  
  
  
  
  
  
own in 
ble 5. 
photo 
45. min. 
30 min. 
60 min. 
30 min. 
2 h.45 min. 
  
——————————— 
photo can 
[S case, due 
to line 
3 operator 
long the 
use boundary, 
je for the 
of a 
work a 
le line. 
f buildings 
€ floors. 
tion east-west 
ave been 
ave been re- 
s known, the 
ned from the 
y land use 
category. The time spent on the data collection and the 
results obtained are shown in table 4 und 5. 
Table 4: 
Time spent on sampling along traverses with a stereoplotter 
  
  
  
  
orientation of photos à + 2 h 
in stereoplotter : 
air photo interoretation 
E : z 3 h. 
and scanning of traverses 
digitizing boundary of : 
test area 30 min. 
TOTAL 4 h.30 min.-5 h.30 min. 
  
  
  
5. Conclusions 
The results of the three different sample surveys have been 
compared with the results of a recently accomplished cadastral 
survey which served as a reference (compare table 5). It can 
be seen that the three methods are equal what their accuracy 
is concerned. The error estimations show a rather good 
conformity with the actually obtained results. 
Conventional air photo interpretation combined with sampling 
techniques have proved to be an excellent tool for obtaining 
statistical data on land use. Among the three procedures, the 
analytical point transfer followed by the point by point 
interpretation has turned out to be a very efficient approach 
for a countrywide land use inventory. 
The amount of time spent for the data collection allows a 
first estimation of time and costs needed for a countrywide 
survey. As they have been obtained under experimental 
conditions, they still will have to be confirmed by more 
extensive studies. 
Acknowledgments 
The presented investigations were carried out in close coope- 
ration with Dr. O. Kälbl from the Swiss Federal Institute of 
Forestry Research in Birmensdorf.