The introduction of artificial parallaxes to produce stereomates is
only possible in the direction of the profiles (i.e. forward or
backward), therefore the direction of the profiles should agree with
the flight direction (within a few grades).
The geometric correction for regularly distorted images is acm
present only by affine transformation. In future it will be extended
up to a third degree polynomial correction.
Operational considerations
A standard OR-1 (also the ITC version with the present program
"ORI") requires input from a magnetic tape. It consists of an iden-
tification number, the scan length, the slit width, the photocoord-
inates of three orientation points and the "image profiles". "Image
profiles" are a list of photo coordinates of those points which
should be imaged at 1 mm intervals along the scan lines in the
orthophoto.
Such tapes can be produced directly by scanning stereo models in a
photogrammetric plotter which is specially equipped for this purpose
with a tracking system and a recording device for photo coordinates
(e.g. WILD PE3 + PBKIl [ 4PAZI ]). The recorded data can only be used
for the production of an orthophoto with fixed scale and fixed slit
size from one specific photograph. Stereomates can not be made with
. this system. The area must be re-scanned to produce an orthophoto
from a different photograph or with a different slit size or at a
different scale.
Another way to produce the required input tapes is to use a digital
computer to convert any height information into the necessary "image
profiles". There are programs available for the production of ortho-
photos, stereomates and other kinds of geometric image transforma-
tions, e.g. the SORA programs [31.14]; To find the transformation
parameters these programs require (amongst other data) photo coordi-
nates of control points. These coordinates can be measured directly
in the new Wild version of the OR-1. Subsequently the data must be
transferred to a computer provided with the SORA programs. To be
able to abandon the empirical orientation of the photograph, the ORI
must be left idle until the SORA processing is ready and the
magnetic tape is transferred. Then the production on the OR-1 can
start.
There is no time lack between measurement of the control points on
the OR-1 and production of the orthophoto (or other product) in the
ITC approach. The measurement of up to 10 points does not take
longer than an empirical orientation. This is because prepositioning
is done by the computer after the measurement of the first two
points. Only very small manual corrections are required after the
third or fourth point if there were no blunders in the first
measurements. Blunders can be eliminated interactively.
For perspective, affine or polynomial image transformation the
computations are very easy and therefore an external computer is not
needed. All the system needs is a ground control file, measurements
196