al
om
Data Collection for Bundle Block Adjustment
on Analytical Plotters
Karsten Jacobsen
University Hannover
FR of Germany
Commission III
Introduction
Program for Data Collection
Experiences,in Data Collection
Experiences with Block Adjustments
Need of Online Triangulation?
Cn .» co P5 +—
1. Introduction
Up to two years ago, the photo measurements for bundle block adjustment of
the University Hannover have been done with comparators. The measurements
have been time consuming and no online check was possible. Caused by this,
a photo arrangement of 60 % endlap and 20 % to 35 % sidelap was used and a
lot of blunders have been included in the data. In addition, the reliabili-
ty of the computed object point coordinates was poor, because of the small
number of photos for each object point.
Now the use of modern analytical plotters has changed the data coilection
and this also has influenced the block constellation. If a point was
measured in a photo or model, this information can be used for an automa-
tic accurate or approximate pointing in other photos or models. So the
use of photos with a higher overlap will raise the time for data collection
not so much.
The topic of this paper is only the data collection for high accurate
bundle block adjustments. That means, the object points are signalised or
they are well defined topographic objects.
2. Program for Data Collection
The advantage of analytical plotters is depending upon the software. Usual
the standard software is limited in the possibilities. By this reason, a
program for supported measurement of photo coordinates for bundle block
adjustment was made for the Zeiss Planicomp with the computer HP 1000.
This program has following main possibilities:
- measurement of fiducial marks or reseau points with automatic approxi-
mate pointing;
- measurement of photo points, partially with approximate or accurate auto-
matic pointing;
- optional second measurement of the photo points in reverse sequence;
- second measurement of fiducial marks or reseau points;
- computation of relative orientation including statistical tests and
transformation of the model to the approximate computed block.
Only numbers of points which have not been measured before or for which no
approximate position is available, have to be typed.
The fiducial marks or the reseau points are numbered corresponding to
their location in the camera. That means, the same fiducial mark can be
sometimes on the left hand side and sometimes on the right hand side in the
analytical plotter, depending upon the flight direction of the individual
strip. The automatic numbering will be done corresponding to the answer to