le 
jue 
ser 
en- 
ais 
he. 
par 
jue 
des 
ble 
rop 
to- 
pa- 
Or- 
10n 
‘on 
ant 
ans 
the 
per 
ree 
ble 
of 
ved 
rial 
ra 
be 
hic 
of 
'to- 
  
  
LACUNES DE LA RESTITUTION, DISCUSSION 181 
graphers to the stereo model and the ease of 
plotting the detail was varied. I vividly remem- 
ber one discussion between an office photo- 
grammetrist and a field topographer. “This 
looks very good but how do you draw contours 
in this dense vegetation?” asked the field man 
sarcastically. “It is very simple, we guess at 
them in the areas of dense vegetation” said the 
photogrammetrist. 
Several decades later our equipment has be- 
come rather more sophisticated, but we still face 
the problem of how much guessing should a 
photogrammetrist do in stereoplotting. We 
should not forget that it is the human operator 
who puts the art into photogrammetry, for it is 
he who interprets the photographs and translates 
the data into useful and meaningful cartog- 
raphic products. It is the map — the end pro- 
duct — which justifies the existence of photo- 
grammetry. The value of photographic maps — 
regardless of how they are compiled — is eval- 
uated in terms of how completely it meets the 
requirements of the user. As Dr Le Divelec 
notes in paragraph 1.7 of his paper: perfect 
accuracy is not a possible human achievement. 
Consequently, we must accept the fact that the 
degree of accuracy and completeness we can 
expect to be portrayed will be limited by the 
nature of the photogrammetric techniques and 
the ability of the human operator. I am firmly 
of the opinion that the role of the human 
operator is the most important link in photo- 
grammetric procedures. I propose that it is the 
most important factor we should consider when 
we discuss the ways to overcome “omissions, 
errors and photogrammetric plots. I cannot 
accept the philosophy that voluntary gaps 
should be accepted as part of a less troublesome 
solution, and that the plotting of a feature which 
is hard to discern leads to mistakes and should 
be stopped. 
We know from the mathematical theory of 
statistical decisions and the probability of 
making a correct decision, whether it be under 
certainty or uncertainty, is not a simple opera- 
tion, but neither is it so complex that it cannot be 
solved by the stereo operator when necessary. 
His development in this respect can be calibrat- 
ed only if he is expected and required to make 
decisions. If we reduce his task to a mere plot- 
ting of sure features and allow him to stop the 
plotting in areas which are difficult to interpret, 
| am afraid that we shall develop a generation of 
photogrammetric compilers who will be hob- 
bling on the crutches of the pre- and post-editing 
rather than inheriting the intellectual power to 
solve the photogrammetric plotting problems. 
Perhaps my feeling on the place of the human 
operator and what should be expected from him 
is influenced by my close association with the 
field of photogrammetric mapping to support 
the photogrammetric charting of constructions. 
If we accept the limitation of aerial photography 
and limit our photogrammetric plotting to fea- 
tures which are definite, then it can contribute 
very little support to surveying. We expect our 
compilers to make decisions, such as where is the 
location of the high water line; where it is possi- 
ble we supply him with photographs taken at high 
water stages and simplify his task considerably. 
Where are the shoal areas which are dangerous 
to navigation, such as sunken rocks and so 
forth? What is the classification of the foreshore 
and backshore areas? These are not features 
which one can definitely see on a photograph or 
as a definite image. To get these answers by 
field pre-editing is a time-consuming task and 
not resorted to except in the most complex areas. 
Photo-interpreter techniques are the most useful 
tools. We attempt to give the stereo compiler all 
the information which is available for him to 
make his decision and accept the fact that there 
may be gaps in the plot. In this connection, we 
feel that the aerial negative is the most important 
source of information, and we concentrate upon 
the photographic processing techniques which 
enable him to extract the maximum information 
content from the aerial negatives. For example, 
prints and diapositives and plotting plates are 
prepared with automatic electronic equipment 
under strict photographic laboratory control. 
In the final analysis we must consider the 
total effort, that is the cost to prepare the map; 
the efficiency is not increased if the effort is 
shifted from the plotter to the field topographer. 
Hindsight is wonderful but pre-checking is not 
efficient unless we assume that the operator will 
miss or mis-identify or omit certain features. 
Photo-interpretation is extremely efficient. 
We should collect and train stereo-plotters with 
the ability and capacity to become good photo- 
topographic interpreters. Let us not subscribe to 
the concept that the topographic maps can be 
efficiently prepared by semi-skilled stereo- 
operators. A good-quality map in addition to 
accuracy needs topographic expression. It is a 
skilled art which must be cultivated. 
In conclusion, I might mention that photo- 
grammetry is one of the many tools which can 
be used on the construction of a map. Its use is 
always justified when it is the most efficient tool 
to be used. As in everything else in the world, 
we must always accept a risk. The amount of 
risk which can be tolerated is a matter of per-