33 
proficiency levels of interpreters, and to the integration of airborne data with 
data from other sources. The department runs a 9 months standard course, 
and an advanced course, in the use of aerial photography for urban planning, 
administration and management. In a special exercise programme, several 
interpretation techniques are applied, using a variety of instruments, photo 
types and photo scales; these exercises also give examples of the data that can 
be collected from aerial photography and the methods, techniques and 
materials that should be used. 
With respect to the level of proficiency, the interpretation activities can be 
split up into several classes (Hofstee 1973). 
Class one mainly concerns the detection, recognition and identification 
of objects, the counting of objects, measurements of objects, areas, slopes and 
heights, the computation of ratios and detection of changes over a period of 
time. These activities can be carried out, after brief instruction, by every 
interpreter; no specific professional background knowledge is required. The 
second class refers to photo-interpretation techniques, which involve a certain 
level of knowledge which enables the interpreter to develop, in co-operation 
with the instructor, the classifications to be used, and to assist in the pro- 
cessing, analysis and presentation of data. He must be able to make the 
correct decisions in problems of classification. 
The third class is represented by specialistic knowledge in a certain discipline 
combined with considerable experience in the field of photo interpretation. 
The specialist in this field must be able to organise, supervise and evaluate 
photo-interpretation projects, and he must be qualified to train interpreters 
of first and second class level for their specific jobs. 
Automatic Processing of Imagery 
“Classification of MSS data by computer permits aerial analysis on an 
unprecedented scale with a minimum expenditure of time. Also, classification 
results using automatic data-processing procedures are consistent, comparable, 
and replicable" (Todd et al 1973). 
Aside from the validity of this statement, a number of questions are not being 
satisfactorily answered by current research. A minimum expenditure of time 
can be seen only as part of a more complex problem of a minimum of time 
and cost to obtain the required results. In an automatic classification pro-