—
=e = Block
— - 2Strips
lE oo
a pacis
PS Used
ol Strategies.
tres).
| this problem,
Section 6) to
rip, identical to
on by half a
is will improve
sults for the full
ontrol as in the
recision. When
nent, it can be
when only four
s of the block.
je for reduction
1easurement is
g that reducing
rall cost of a
The precision of the antenna coordinates was derived
from a comparison with the true values. Unfortunately,
these simulated coordinates were of an unrealistic
precision. The simulated data was immune to multipath,
geometric effects and cycle slips that would be expected
in real flight, giving the results limited physical meaning.
The analysis was extended to investigate how the ground
point precision was affected by gradually reducing the
precision of the antenna coordinates (between the range
1cm - 5cm at 0.5 cm intervals). This showed results from
using more realistic antenna coordinate precisions. Three
control point configurations were used, namely no control
points, 4 full 3D control points in the corners of the block,
and 4 full control plus 6 height control points within the
block.
Height SE (mm)
No GPS Used 15 25 35 45
GPS Precision (cm)
Figure 4 - Effect Of Antenna Precision Degredation On
Unknown Point Height Precision (Three Control
Strategies).
Figure 4 shows how the antenna precision effects the
overall height precision of the unknown points measured
in the block. The same data was used throughout the
analysis under the three afore mentioned control
configurations, with only the antenna coordinate/SE file
being altered for each estimation. It is clear that the
degredation causes a near-linear degredation of precision
when no ground control is used within the photographic
block. By comparison, the controlled blocks do not show
such a marked degredation. It is not clear from Figure 4
what the relationship is between the other two
configurations. To investigate this further we must look at
Figure 5a) and b) where the no control strategy has not
been plotted.
As would be expected, the configuration with the least
height control is most susceptable to antenna precision
degredation. When height control is used throughout, the
precision degredation of the heights of the unknown
points is minimal between the range 2cm - 5cm. Even at
5cm antenna precision, the contribution to the estimation
process is still a positive one. The configuration using
only 4 Full 3D control points shows a clear degredation by
comparison. However, this is still a sub-millimetric
difference between each 0.5cm antenna precision
interval.
Figure 5 shows that the height precision of the control
points degrades in a similar manner under both
configurations. The use of additional height control
appears only to transpose the line to a higher precision
level.
6. Further Combined Estimation Work
The simulations discussed here have concentrated on the
high precision end of antenna coordinates. This work is
now being extended to investigate the properties of this
degredation from 5cm precision down to 15cm precision.
Also, in addition to these estimations, work is now
progressing into simulations of different flying heights,
flying strategies and control configurations for a longer
photographic flight. The data for this work is being
completely simulated, using knowledge of the real
properties of photogrammetric data at this scale. The
GPS data simulation can be improved in a way which will
allow the investigation of how satellite geometry affects
the final bundle coordinates and standard errors.
je = 4 Full 3D + Heig!
Height SE (mm)
a
o
a uices» messe, emer
pu d wem co TO
-- *
ne
No GPS Used 15 25 35 45
GPS Precision (cm)
\ Pot i i i 2 E [== =4 Full 3D
(m p nt
Helght SE (mm)
[— —^FuisD ^ |
|—0 = 4 Full 3D + Height
No GPS Used 15 25 35 45
GPS Precision (cm)
Figure 5 - Effect of GPS Antenna Coordinate Precision Degredation On (a) Unknown (b) Control Point Height SE Under
Two Different Control Strategies.
819
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996
