6. How correct are the individual interpretations? 
In order to be able to answer this question, a 'standard interpretation' 
was made. This standard interpretation differed from the other inter- 
pretation in several aspects: 
— the interpreter has considerable experience in the interpretation 
of urban airphotos; 
— the interpreter is highly motivated to make the best possible 
interpretation for research purposes; 
— the interpreter is familiar with the city and its building 
styles through field work; 
— the amount of time spent on the interpretation is much longer 
(six hours against three hours for the normal interpretation); 
- the interpretation has the character of a check: all areas are 
analysed in detail with the sieve map as guide. 
The standard interpretation shows 48 ha as newly built-up. When the 
individual interpretation is compared with the standard interpretation, 
three categories of areas can bé distinguished: 
— areas correctly interpreted as newly built-up; 
— newly built-up areas 'overlooked' by the interpreter; 
— areas 'incorrectly classified' as newly built-up. 
After checking each interpretation, the areas were measured with a 
dot-grid (9 dots/ha). 
The areas correctly interpreted plus the. areas 'overlooked' should 
always be the same. It proved to be 48 ha + 2% (standard deviation 
1%). (This is a test on the correct checking of the interpretations). 
The area correctly classified by each interpreter differs fyom 87% 
to 96%, with an average of 92%. (See table 2, column 1). This means, 
of course, that the area 'overlooked' differed in size from 4% to 
13%, with an average of 8%. (See table 2, column 2). 
The areas overlooked are spread over the whole area, There is no 
difference between foreground and background areas on the obliques. 
The figures 4B and 4C show the overlooked areas. They tend to cluster 
near the edges of the obliques, which might be explained as orientation 
difficulties. Figure 4B shows the 15 areas overlooked by one to five 
interpreters, while figure 4C gives the six (smaller) areas overlooked 
by more than five interpreters. 
The areas incorrectly classified as newly built-up are shown in 
figure 4A. They are somewhat smaller than the areas 'overlooked'. 
Co: 
In 
Av 
hi 
TA