113
8
though we would like to have it more strict than we could make, it has worked
rather well. At least, these specifications have had a uniforming effect in the meth
ods of camera and lens testing. I should like to make a few amendments to the
existing specifications. This draft for testing methods supplies information about
two groups of data; one group is »definition of the qualities of the image», and the
other group of data defines the »geometrical accuracy». As to the first group
of information, the image quality, this is at the moment defined by means of the
resolution. We know that the justification is not a physical one but it is a purely
pragmatic necessity to use this measure as a means of comparison, and to
standardize all circumstances when we determine this resolution. We also know
that at the moment a better insight in the phenomenon of image information,
image sharpness, is being gained. And we hope that in a not too distant future
image sharpness can be defined in a stricter way than it is possible today, that
it can be well defined in plain physical terms. I therefore propose that we leave
the paragraphs regarding resolution unchanged for the time being. They will
have to be changed, and be brought uptodate at some future occasion. The
paragraphs regarding image illumination, shutter efficiency and veiling glare
provide good standard methods and they fulfil also future needs.
Now we come to the second group of measurements defining geometrical
accuracy, i.e., the determination of the elements of inner orientation of a photo-
grammetric camera. I have two amendments to these measurements. The first
is the controversy between photographic methods and visual calibration
methods. So far, many institutes are still using visual methods, other institutes
use photographic methods. Having modern cameras of very high performance,
the methods of calibration should be brought up to the highest possible
standards. There are indications that the photographic method and the visual
methods do not give identical results so; it is necessary to clarify the situation
and if necessary, in the near future, to decide upon the purely photographic
method as the standard method for camera calibration. The second amendment
is the use of the point of symmetry. We have calibrated a number of different
cameras made by different manufacturers from different countries. These results
show that in a non-perfect lens, in an actual lens, the principal point of auto-
collimation and the fiducial centre and the point of symmetry do not coincide.
In the cameras we have tested, the differences between these points are of the
order of 30, 40 microns and sometimes 0.1 and 0.2 mm. These values cannot
be neglected because they result in appreciable assymetric distortion. Now it is
a sound practice of many institutes that they determine the position of this
point of symmetry, but it is confusing if the resulting point is then called princi
pal point, as it is sometimes in practice. Therefore I propose to define the »point
of symmetry» also in this proposal,—the best way is indicated by Roelofs in
his wellknown paper in »Photogrammetria»,—and then to include this determi
nation in the draft. This has several important advantages for practical opera
tions.
The last amendment is, that there is still a confusing situation regarding the
definitions which we use. Even the most basic concepts such as principal point,
principal distance, are not yet internationally standardized. It is not only the
task of commission I, but also of other commissions to agree and to promote
international agreements on these basic definitions. This is urgently necessary.
Conclusions: First of all, the paragraphs relating to the determination of image
quality should be left as they are.
It should be clarified whether photographic calibration is necessary and if
so the photographic method should be explicitly specified, the point of symme
try definition should be included and its establishment should be specified and
definitions should be internationally agreed upon.
Mr. F. L. Corten se declare assez satisfait des essais du projet de norme. Il espère, comme le
Dr. David, que dans un avenir rapproché on pourra donner une définition plus stricte de la qualité
de l'image et propose de laisser en l’êtat la question de la résolution pour l’instant. Il trouve