5.0 PERFORMANCE WHEN RESET
We had hoped to use the output from a block adjustment to reset the MPS
and simulate the situation where an interior orientation only is performed
and the absolute orientation is read in from a diskette. However, no
adjustment data was available for the particular stereopair used in this
test so the orientations from the space resection/intersection were used
instead.
Several factors have affected the accuracy:
I. There was difficulty in indentifying the control points in certain
positions. Points 31, 43 and 50 are probably biunders.
2. The floating mark was too large when the MPS was used at maximum
magnification. For this type of work a smaller mark should be used.
3. There seemed to be some non linear shrinkage as indicated by the
measurements on the fiducial marks. This would have a significant
effect on the observations and only further measurements will settle
the source of this trouble.
The biggest problem was the difficulty in finding the exact points used
for control. There had been changes such as road resurfacing in the town
and some points were on the ground near buildings.
TABLE 1: Times for This Exercise
Mounting diaposatives 2 minutes
Interior orientation 5 minutes
Exterior orientation Not needed
Observing data points 30 minutes
6.0 MEASUREMENT WITH ANALOGUE INSTRUMENT
The original large format diapositives were set up in a Kern PG-2
stereoplotter fitted with a tri-axis locator and digital output. Setting
up was carried out by a practising photogrammetrist, using datapoints
which had been checked by the previous tests. There was some difficulty
in setting up the photographs due to the large overlap and for this reason
a relative orientation oniy was performed and the model co-ordinates as
measured were then transformed to make a best fit with ten selected
control points.
TABLE 2: Times for This Exercise
Mounting diaposatives 10 minutes
Interior orientation 80 minutes
Exterior orientation Not performed
Observing data points 30 minutes
26
