134 GRAPHICAL OR NUMERICAL PHOTOGRAMMETRY?, DISCUSSION 
of the results separately achievable from graph- 
ical and numerical photogrammetry because he 
has made those points very clearly in his paper. 
However, it would be inappropriate if I did not 
mention the fact that in engineering especially 
numerical and graphical photogrammetry go 
hand in hand, and so it could not be a case of 
graphical or numerical. It is really a case of 
graphical and numerical photogrammetry being 
utilised throughout the separate stages of en- 
gineering. 
We can think about numerical and graphical 
photogrammetry as separate subjects. We must 
also recognise the fact that they are dependent 
strictly on the recognition and interpretation of 
the features which shall be measured, or the 
features that we desire to measure, and therefore 
it revolves around the ability of individuals and 
the information which they desire to get from 
the photographs, a choice of one versus another, 
strictly from a utility point of view. 
When graphical photogrammetry is utilised, 
most of the results can be obtained stereoscopic- 
ally to a better degree of accuracy and fuller 
scope. However, sometimes non-stereoscopic 
photogrammetry can be applied. It can be 
applied more fully in the numerical than in the 
graphical. The graphical results of photogram- 
metry are always required in engineering regard- 
ing traverses and profiles and cross-sections and 
the plotting of cadastral data for rights of way. 
The graphical may have come, of course, from 
the numerical, and in many cases the reverse 
is true. We first get the graphical and then we 
convert it to the numerical. 
The degree of accuracy required in each one 
of the uses of these two types of photogramme- 
try will depend on the stage at which we are 
working. We have reconnaissance stages in 
which the accuracy can be low because we do 
not require precision in reconnaissance stages. 
We are more interested in differences rather 
than specific accuracies of points and detail. As 
we get further along in the steps of our engi- 
neering the accuracy increases and the scope 
of detail decreases, and we get to the point 
finally where we need to have precision to the 
nearest fraction of a metre and sometimes centi- 
metre. I do not believe that we need to go to 
millimetres. 
My time is up so I will just conclude by 
saying that the graphical data and the numerical 
data in engineering are actually inseparable and 
that we utilise them at all stages of engineering. 
It is not strictly a matter of direct selection of 
one as against the other. It is strictly a matter of 
correlation and utilisation as the work proceeds. 
Mr Van der Weele has introduced a very 
excellent subject. 
Mr C. O. TERNRYD: Under point IV in his 
invited paper Mr Van der Weele deals with the 
problem of graphical and numerical photogram- 
metry in civil engineering, and after some dis- 
cussion he makes the statement that a combina- 
tion may be the most suitable. He discusses the 
problem of using photogrammetry in the de- 
tailed phase of civil engineering work. The 
reason for that is that there are so many differ- 
ent means used for the preliminary planning 
that no general procedure exists. 
In our opinion and from our experience there 
is a very great advantage in using photogram- 
metry in civil engineering needs also in the pre- 
liminary planning phase when it is a question of 
finding the best preliminary alignment of a road 
or a water power station. Therefore, the ques- 
tion of graphical or numerical photogrammetry 
also should include this phase for giving a com- 
plete discussion of the question. 
The two greatest problems of using photo- 
grammetry in civil engineering are, first, to 
evaluate the procedure which includes as well as 
the preliminary phase the detailed phase, and 
second — which is more important — to educate 
the engineers to using the picture according to 
that procedure. The procedure evaluated should 
be built up so that the demand on a large survey 
in the preliminary phase can be combined with 
the demand of high accuracy in the results of 
the design phase. 
[ should like to give some short comments on 
the preliminary planning phase as comments on 
the design phase will be given by Dr Zarzycki. In 
this phase, the preliminary planning phase, a 
very great part of the problem can be solved by 
performing the planning directly in the photo- 
grammetric model itself without the necessary 
mapping. This type of photogrammetry is some- 
thing between graphical and numerical. The 
main question is then to find the photogramme- 
tric instrument that at the same time gives a 
good survey over wide areas and a sufficient 
accuracy in the method of numerical data. The 
advantages of using the entire model lies therein, 
that it gives a much greater amount of informa- 
tion than a very detailed map can give you. 
When planning roads, for instance, the model 
gives the information as well as the topography 
of the landscape as well as the geology and the 
drainage of it. My paper presented to this Con- 
ference has been distributed and it gives the 
method of using the Kelsh plotter for the preli- 
minary planning phase. That procedure gives 
  
us € 
getti 
late: 
just 
we 1 
the 
the 
usin 
han 
very 
loca 
It i 
engi 
beir 
will 
edu 
mal 
ists 
aris 
feel 
nin 
fiel 
nee 
ject 
SOIT 
in € 
try 
Wol 
whi 
con 
are 
me: 
mei 
ele 
of 
nur 
sim 
phe 
tak 
des