yistering
ven dis-
lation is
ances in
t angles,
realized.
e deter-
may be
used:
relative
ero dis-
:et-side
erefore
into a
image
osition
mation
elation
data:
on.
ection
which,
b) Explanation of some definitions.
Principal point: The best rewarding problems of standardization, but also the most
difficult, arise in connection with the principal point. The purpose of the principa!
point is that of representing in the image plane an origin of coordinates with certain
properties. For this purpose, the following definitions — among others — are available:
Principal point math. definition
Fiducial center
Optical center
Collimated principal point
Point of symmetry )
phys. definitions
Registering fiducial center
Registering optical center | techn. definitions
Registering collimated principal point
Calibrated principal point
This list can be extended without difficuity. One must question, however, whether it is
unavoidable to burden the terminology with a ballast of multiple designations and
definitions for one and the same thing; for as a rule, all of these points coincide with
the practically necessary accuracy in a single point, while the differences and fine details
are required only in exceptional cases and for special investigations. This is where
standardization should begin. Its purpose cannot be to conserve all of these definitions
in a standards sheet, but must rather be to extract that which is absolutely necessary in
practice and simplify it as much as is at all possible.
Unfortunately, these efforts are pretty closely limited for historical as well as prac-
tical reasons. — According to the above fundamental statements, the technical defini-
tions should be excluded from standardization. Thus the calibrated principal point, which
is obtained by optimal compensation of the asymmetry errors [2], will be left uncon-
sidered. It goes without saying that this statement is not intended to suggest anything
to the disadvantage of this method. — The points relating to the registering plane may
likewise be counted among the technical concepts and accordingly omitted in stand-
t they are of practical importance only when the image plane and
in the ease of a manufacturing error. Apply-
1 differentiation between the image plane and
the registering plane also to the corresponding points will be superfluous in most cases.
— The fiducial center given by the fiducial marks, though as a rule coinciding with the
principal point, nevertheless has an independent conceptual significance. Neither car
it be denied that also the optical center (point of penetration of the optical axis through
the image plane) has its raison d'étre, the more so since, if the Scheimpflug condition
is realized, this point is at a finite distance from the prineipal point (the optical center
is not mentioned in the Table because it constitutes not a photogrammetrie, but ar
optical concept). — There remain the three definitions for the principal point: The
mathematical prineipal point, the collimated principal point, and the point of symmetry.
No question can exist that the mathematical definition is indispensable. The two physical
definitions differ by the measuring method. The collimated principal point is obtained
by autocollimation, ie. by placing the principal object axis perpendicular to the imag
plane. The point of symmetry, on the other hand, is won by resection in space [2]. Both
methods are in continual use. For the physical definition of the principal point, however
one or the other may be dispensed with without drawback; for since freedom from
manufacturing errors is presumed, the two methods will necessarily furnish exactly the
same point. As has been shown elsewhere, the method of autocollimation for deter-
mining the principal point is far superior to that of resection in space, and especially
ardization, seeing tha
the registering plane do not coincide, i.e.
ing the frequently useful purely conceptua
11
