Software Wedge for Windows from TAL Enterprises, but
similar software from other vendors may work as well.
2.3 PC-TAS Excel sheet
The PC-TAS Excel sheet is divided into several parts:
input/output part, calculation part and a results sheet.
Normally, the students will only notice the input/output
and the results sheet parts unless they really want to
study the way the calculations are done.
Excel workbook
Sheet 1
It is possible to gather all calculations and formulas on a
sheet of its own, effectively linking it to both an
input/output sheet and the results sheet. However, tests
have shown that this will slow down the program
significantly. The current version of PC-TAS thus have
an Excel sheet in common for input/output and
calculations, as can be seen in the illustration above.
The PC-TAS input/output part is in turn divided into five
subsections. The Excel sheet is too large for a total
screen view, so the user will have to scroll from one
section to the next. The subsections are:
1. Inner orientation/Transformation
2. Object measurement
3. Relative orientation
4. Absolute orientation
5. Object coordinates
Interactive interface. Unlike most Excel sheets where
calculations are automatically performed immediately
after the input, PC-TAS has an interactive user interface
in the meaning that the user must click at indicated
"buttons" to view the results from the different
calculations. The purpose of this is to allow the students
enough time to follow and understand the calculations
and to check the intermediate results.
112
3. USING PC-TAS
3.1 Preparations
In order to run PC-TAS, the user first has to connect the
digitising table to the computer and set the configurations
in Software Wedge (or similar software) to match the
settings of the digitising table. These configurations
includes baud rate, input port, parity, number of data
bits, input string length etc. This is done only once, i.e.
when the Software Wedge is installed. The settings are
stored for future use.
To set up a stereo model, the user places a pair of paper
print stereo images on the digitising table, making sure
that as many fiducial marks as possible will fit inside the
digitising area. The user will now continue to work with
the PC-TAS Excel sheet, as will be explained below.
3.2 Operation
Section 1 - inner orientation. This establishes the
affine transformation parameters for the transformation
from digitiser coordinates (xD yD to image coordinates
(x' y', x" y^). Since PC-TAS uses a two-dimensional affine
transformation to determine the transformation
parameters, at least three fiducial marks must be
digitised in each image, and they must not be in a
straight line. The user will digitise the fiducial marks into
the Excel sheet in their respective cells by activating the
correct cell and digitise. On the users command
(triggered by clicking at a button), Excel will calculate the
affine transformation parameters. These parameters will
be used from here on to transform all indata to the image
coordinate systems. Presently no corrections are made
for radial distortion or atmospheric refraction, but it would
be easy to include them at this stage of computation.
If the solution of the transformation is not satisfactory,
the user can easily re-digitise some (or all) of the
fiducials. To help in this judgement, PC-TAS also
displays the residuals from the comparison of known
fiducial coordinates and the transformed digitised
coordinates.
Section 2 - Object measurement. In this section the
user will digitise the objects he/she wants to measure. In
its current form, PC-TAS allows for 25 such object points
to be digitised.
Digitisation is performed as mono-measurement
separately for left and right image. It is possible to equip
the digitising table with a stereoscope, thus enabling
stereo vision while digitising. The measurement is,
however, done in one image at a time as there is only
one cursor.
The object measurement is followed by computation of
model coordinates, using the normal case parallax
equations.
The reason that this step precedes the relative
orientation is to facilitate a later comparison of model
coordinates (x y z) and parallaxes (px, py) before and
after the relative orientation has taken place, in order to
give the students some understanding of the importance
of relative orientation.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B6. Vienna 1996
Sei
will
car
dig
the
Thi
use
The
dis
cor
rele
Se
ent
ide
per
TA
dig!
exc
will
int
-H
- Le
Se
but
obj
coc
Re:
ane
Re:
err
eve
