198 PRESENTATION OF THE GENERAL REPORT OF SUB COMM IV-4
particular the phi element. This limitation had to
be emphasised and never forgotten when apply-
ing methods where we measure differences or
parallaxes. We should not regard the parallax
method as a major means which solves all
problems, but we should take it for what it is: a
convenient and inexpensive aid for checking the
fundamental óperation (3) of photogrammetry,
the relative orientation.
Under normal circumstances, in particular
with respect to the first two operations, we can
get very valuable information about some
systematic or regular error of the combined
operation (1) and (2), and primarily about the
irregular errors. The control experiments have
to give important qualities satisfactorily. We can
doubtless find certain specifications upon such
measurements, but of course we shall always be
grateful for checks from extra control points in
the models.
Evidently, no detailed analysis of the individ-
ual co-ordinate errors can be performed from
such measurements. It is more in the statistical
sense that we use such measurements here.
Finally, I should like to emphasise that the
work and results of Sub-Commission IV-4 are
nothing but a check of the method of least
squares, law of error propagation and the cen-
tral limit theorem from statistics. It may also be
mentioned in this connection that the principles
which have been applied to the work of Sub-
Commission IV-4 are by no means a recent in-
vention. The first test, according to this method,
was made in 1937—1938 in connection with the
introduction of the Multiplex instrument to
Sweden. In April 1938 a proposal for tests of
this type was written up and forwarded to the
Geographical Survey Office of Sweden. In 1944
the basic principles were published.
After these introductory remarks Prof Hallert
showed a number of slides representing the
following tables and diagrams of his report:
1. Table 1, p 6 gives some information about
the accuracy obtained with Wild Autograph A7
and Zeiss Aereoplanigraph C8.
2. Table 1 of the "additional remarks" to the
General Report IV-4 shows a detailed compa-
rison between the theoretical determination of
the accuracy from y-parallaxes which we have
here, and the corresponding determination from
errors on the ground between geodetic and
photogrammetric co-ordinates. He stresses the
agreement between both values.
3. Table 2 of the same document gives the
same comparison for the results of the Renfrew
test of Commission IV-3.
4 and 5. Histograms p 14 of the General
Report IV-4.
6. Fig 7a and 7b p 18 of the General Report
IV-4. This summary of the distortion deter-
mined by Hallert from observations of a number
of participants proves that the standard error of
the average is in accordance with what we can
expect theoretically.
Then Mr Hallert continues as follows:
Anyhow, I think we can state that these ex-
periments have given us astonishingly good
results which may be of use for future develop-
ments in photogrammetry.
Dr H. HArry: Thank you very much. Yo
will have an opportunity to discuss this com-
munication on Monday at 15.40 in room 381.
E o es