Photogrammetria, XIX, No. 6
in the not too distant future instruments containing one or more of these devices will be
in normal operation.
In order to evaluate the performance of these modern instruments a new approach
will be required. For the present instruments it has been normal practice to restitute a
set of gridplates and compute the accuracy of this restitution but it will be obvious that
such a result characterises only a part of the performance of the instrument.
Although the modern development will increase the flexibility of the system, it also
increases the probabilities of mis-adjustments, failure and trouble and thus may require
more frequent tests and of a different type than has been usual up to now.
These considerations tend to the application of the information theory (and system
analysis) to describe the performances of a photogrammetric restitution system since
this approach and its terminology enables us to describe in a consistent and systematic
way and with adequate words, the phenomena in which we are interested.
We do not intend to develop this idea completely in the present paper, but we will
use a few of the concepts of the information theory and system analyses to describe the
performances of what we are inclined to call the classical photogrammetric instruments,
in which the photographs are considered as the ‘input’ information and the map or the
coordinates of points are the ‘output’.
I. 4. The quality of the ‘output’ depends in the first place on the ‘input’, or in other
words on the quality of the image which is presented to the operator (or to the
automatic scanning device). In this paper we will not deal with the quality of the photo
graphs, but we are mainly interested in the way in which a given set of photographs
is presented and the amount of information that may be lost on the way to the
operators’ eyes. When we consider the original negatives as the basic source of informa
tion, it will be obvious that we should include in this consideration the process of making
diapositives.
This photographic process can be related to the properties of the photogrammetric
system in different ways e.g. by the necessity to combine direct copying of negatives
with the application of corrections for lens distortion, refraction, earth curvature a.s.o.
or by the necessity to make reductions eventually combined with corrections.
For the existing instruments we can thus say that the optical performance is an impor
tant criterion for their evaluation. We will not, however, enter into details of this subject
here. The processing of negatives is more properly the field of Comm. I, and the informa
tion transfer in the instrument itself is more a criterion which is important for the choice
of an instrument than for its performance as a function of time.
There is, however, one optical property of the instrument which may change with
time and which has an influence on the image quality. This property is flare, due to
reflected light in lenses and prisms, light scattered by dust particles or fingerprints on
mirrors a.s.o., and light entering the path of the information from outside (daylight,
lamps).
All instruments have flare, specially those with complicated optical systems and
this flare will generally increase with time by the appearance of dust, scratches a.s.o.
Flare can be measured with a densitometer and its influence can be decreased by cleaning
the optical parts and protecting the instrument from outside light. A peculiar phenomena
seems to be that most operators are not very sensitive for this influence although their
observations may be essentially affected. In one instrument we measured an amount of
false light of an intensity equal to the darker part of the diapositives with maximum
illumination, without the operator complaining about or even noticing this fact. There
fore, a regular test of the instrument and its surroundings with respect to this effect,
has definite value.
In this paper we will pay special attention to the ‘output’ of the photogrammetrical
