GEOMETRISCHE EIGENSCHAFTEN DES BILDES, AUTHOR'S PRESENTATION 
The irregular errors can naturally not be corrected in analogous instruments, in 
plotters, which operate according to the mechanical or optical projection principle. These 
errors must therefore be as small as possible. 
RESUME. 
Partant de l'acecumulation d'informations dans le négatif en cas de reproduction 
idéale, les déviations du standard (cas idéal) sont traitées, qui proviennent des éléments 
individuels en dedans et en dehors de la chambre au moment de la prise de vue. Ces 
déviations sont de caractères variés. Des erreurs projectives sont des erreurs qui peuvent 
être éliminées exactement ou suffisamment par une définition appropriée des éléments 
de l’orientation intérieure de la chambre ou par la projection dans l’appareil de restiti- 
tion. Des erreurs non-projectives (par exemple distorsion) peuvent être symétriques par 
rapport au centre ou irrégulières (par exemple déformation de l'émulsion). Des dévia- 
tions symétriques par rapport au centre peuvent étre éliminées relativement facilement 
lors de la restitution. Il semble done utile à ce sujet de traiter les éléments de correction 
utilisés comme partie de l'aceeumulation et d'observer la géométrie de l'image aprés une 
telle correction. 
Il est évident que les erreurs irréguliéres ne peuvent pas étre corrigées dans les 
instruments analogues, dans les équipements de restitution, qui travaillent d'aprés le 
principe de projection mécanique ou optique. Ces erreurs doivent done étre aussi petites 
que possible. 
Author's Presentation of the Paper 
Mr Chairman, Ladies and Gentlemen, since 
the paper „Über einige geometrische Eigen- 
schaften des Bildes und die Elemente der in- 
neren Orientierung von photogrammetrischen 
Kameras” has been published in “Photogram- 
metria Congress No. C” I will only make a few 
remarks on the subject. 
The purpose of the photographic picture is 
to record the information received at the mo- 
ment of exposure so that in subsequent plotting 
of the terrain all the relationships can be re- 
constructed. The information on the terrain de- 
tails is carried by a bundle of light rays having 
its projection centre in the camera station. This 
bundle of rays is cut by the plane of the nega- 
tive, and the points of intersection of the rays 
with this plane define the picture details which 
represent the recorded data. 
As a starting point for these investigations, 
we use the ideal case where the rays are straight 
lines passing through the projection centre and 
the photographic emulsion is a true plane, the 
picture plane. The perpendicular distance from 
the projection centre to this plane is the “prin- 
cipal distance” and the foot of the perpendic- 
ular is the “principal point” of the photograph. 
The inner orientation of the ideal camera is 
given by these elements, namely the principal 
point and the principal distance. Now a con- 
siderable difficulty exists in photogrammetry in 
that the means of transfer of the information 
and its recording are not ideal. They show de- 
viations which result in a distortion of the 
record. Nevertheless, we should be able to deter- 
mine accurately the positions of ground points 
from the information given by this record, so 
that we obtain a true scale model. 
To achieve this purpose it is desirable to 
eliminate the errors introduced into the record 
at their source. To come as near as possible to 
the ideal case, knowledge of all deviations is 
necessary in the first place. For this, the cause 
of each error should be found. This would 
enable one to decide whether the cause could 
be avoided or whether the result of the fault 
could be reduced to a negligible quantity. This 
will often be a question of economy. If the cause 
cannot be avoided, then the question to be in- 
vestigated is whether the deviations can be 
eliminated or reduced by correction during pho- 
tography, namely when the information is in- 
troduced as input into the record, or during 
plotting, namely as output from the record. 
When classifying the errors we find first 
axially symmetrical distortion. This can be 
corrected in the plotter. Secondly, asymmetrical 
errors which can be corrected by alteration of 
the inner orientation, and asymmetrical errors 
which cannot be corrected. Thirdly, irregular 
errors which cannot be corrected. 
The positions where error arises are first, 
the camera. Then we have the lens distortion 
and asymmetry, the filter which gives a prism 
effect and non-uniformity and the emulsion 
where we have buckling and non-uniform bend- 
ing of the emulsion carrier and film shrinkage. 
Then outside the camera we have the errors 
coming from camera port glass, atmospheric re- 
—— 
     
   
  
   
  
   
   
    
   
  
  
  
  
  
   
   
   
   
   
    
  
  
  
  
  
  
  
     
   
   
   
   
  
  
  
  
  
  
   
  
  
   
   
   
  
  
  
  
  
  
   
  
   
  
   
   
    
    
fraction a 
can be re, 
when a fl: 
The ca 
from the i 
instance € 
manufactu 
lens, and 
the camer: 
Radial 
is of the o 
but radial 
mechanica 
tically, usi 
by the cor 
the correc 
rection pla 
of a partic 
or minus 2 
as a maxin 
In prac 
tured for a 
tion that th 
of the sam 
occur in m 
5 microns 
correcting 
thus gener: 
These € 
very small 
film shrin 
caused by 
They origir 
has a dece 
distortion c 
of the opti 
a decenteri 
for the inr 
been done : 
rors causec 
blications fi 
and others. 
An attei 
a new defir 
possible in 1 
point, the t 
in the dire 
Such a mo 
à projecting 
projective 
distortion, 
This is the 
to the pictu 
ed for, it re 
and must tl 
lest possibl 
asymmetric;