Before closing the session, the President thanked the Chairman and this Working Group for their efforts which 
had led to an interesting discussion. 
Saturday July 29 1972, 10:45 
Semantic Information Session 
Chairman: A.J. van der Weele 
In opening the session, the Chairman announced that this topic had been selected, since photogrammetrists 
must also concern themselves with the completeness aspects of information presented in maps, and not only 
accuracy. Since the presented papers for this session were related, the Chairman proposed to first have both 
papers introduced, followed by a joint discussion. 
W. Hofmann, in presenting the OEEPE Commission E report of K. Neumaier entitled "An Interpretation 
Experiment", announced that the aim of the investigation was to determine the most favourable flight 
configurations for topographic interpretation on behalf of the revision of 1:50,000 maps. Three types of areas, 
densely populated flat terrain, forested highlands and mountainous terrain, were photographed at scales of 
1:12,500, 1:17,500, 1:25,000 and 1:35,000 - three camera types being used for each scale, namely normal, wide 
and super-wide angle. In total, 118 interpretations were performed by 10 European organizations. 
In spite of the voluminous data available, no definite conclusions could be drawn regarding the most suitable 
choice of camera type or photo scale for the different types of terrain. The following indications can, however, 
be mentioned: 
- on an average, only 67% of all objects were readily identified in the varying photo scales of the different 
types of terrain; 
- the average interpretation times for the 8 sq. km. flat, hilly and mountainous test areas were respectively 
63, 23 and 20 hours; 
- in flat terrain, the wide-angle photography yielded better results than the other two camera types; the 
results also improved with increasing photo scale; 
- no preference could be found for either photo scale or type of camera in both hilly and mountainous 
terrain; 
- since a decrease in the photo scale will determine the limit of interpretability of various objects, it was 
recommended to extend the experiment to include smaller photo scales. 
H. Jerie then summarized his and J.W. Shearer's presented paper, Research Concerning the Semantic 
Information Performance of Aerial Photography". The interpretation process was clarified as involving 
detection, identification, classification and deduction. The importance of considering the semantic information 
performance cannot be over-emphasized, since map specifications require the presentation of certain objects, 
and the planner has to ensure that this will be achieved. Incorrect planning in this respect can be very costly, 
since there is an important trade-off position between photogrammetric interpretation and the complementary 
field survey or completion phase. 
The factors influencing the semantic information performance were then outlined by the speaker as belonging 
to the following three categories, namely: the project parameters, of which the photo scale is the most 
important; the project area; and factors pertaining to the agency executing the project, which mainly involves 
operator experience. He proposed to consider a different approach to the problem, by attempting to define the 
physical parameters which determine the reliability of interpretation. A research programme was put forward 
involving, on the one hand, the statistical determination of the relationship between the physical characteristics 
of the imagery and the reliability of interpretation of certain objects, and, on the other hand, the determination 
of how the properties of the imagery behave in relation to the influencing factors. 
The Chairman opened the discussion by remarking that the topic of interpretation had been discussed at length 
during the Delft Symposium, and that its complexity was particularly well illustrated in the OEEPE report. A 
number of other experiments were reported, namely, that in Sweden significant improvements were obtained 
with photo scales ranging from 1:30,000 to 1:20,000 through to 1:10,000. The use of diapositives instead of 
paper prints also resulted in a slight improvement, but the skill of individual interpreters remained the most 
significant variable. Another experiment reported was that of Commission D of the OEEPE, where an 
extremely high reliability of interpretation, namely 9096, was achieved with 1:5000 orthophotos. However, it 
was also argued that there was a need for more objective research, since pilot research projects such as these 
do not supply results which are universally applicable, as is borne out by the fact that not all countries require 
the same type of information on, say, a 1:50,000 map series.