4
A
Helght Standard Error (mm).
1.5 Fr 4
No Control d 4Full 4 Fuli+Ht GPS + Fuil Control No GPS Used
[=e = Block
| =e = 2 Strips
—e — Centre
+
Helght Standard Error (mm).
No Control 4 Full 4 Fuli+Ht GPS + Full Control No GPS Used
Figure 2 - Effect of Additional Strips On (a) Control (b) Unknown Point Height SE Under Different Control Strategies.
820 840 860 880 900 920
| L | | i
10404
4 a d
8
5 0207 5 5 o
Ë
8
i 1000 of of
x
3 980-| e o eo
5 e
960-4
T T T T T
820 840 860 880 900 920
Locd Y Coordnate (metres)
940 960 980 000 10
L 1 | 1
1040
d ol
o e -1020
d a 1000
e d 080
Ss S
|-960
—— lli 602.
I I T T
940 960 980 300 20
Figure 3 - Heighting Standard Error Distribution Plot for Centre * Lateral Strip (vector unit is metres).
Figure 2 show how the photographic configuration
influences the height standard error for the block when
the control configuration is altered (they are presented as
lines passing through the 5 points defined by the x-axis
labels). The addition of a lateral strip to the main centre
strip (the 2 strips line) gives a marked improvement in
overall precision. The improvement gained by adding the
third strip is not so great, which is encouraging when the
cost of additional strips of photography is considered.
However, with a lateral strip flown in this manner (6096
overlap) much of the photograph is unused in the analysis
and the far set of control points are not covered. This
leads to the standard error pattern for the heights shown
in Figure 3 (only the magnitude of the vectors is
important). All the far points (those on the top row) have
higher height standard errors because they appear on
less photographs. The near points (those on the lower
lines) show an improvement when compared to a similar
Control Config.
GPS Block 24 2.1
GPS 4 full + height 3.0 23
GPS 4 full 3.2 2.5
GPS no control 8.9 4.9
Unknown Point Standard Errors (mm)
plot of only the centre strip. To alleviate this problem,
simulated analysis is underway (see Section 6) to
investigate the use of a second centre strip, identical to
the first but offset in the flight direction by half a
photograph 'length'. It is expected that this will improve
the standard error pattern from Figure 3.
Table 4 gives a detailed summary of the results for the full
block of photography. Processing without control as in the
last row leads to a marked decrease in precision. When
control points are included in the adjustment, it can be
seen that heighting precision is decreased when only four
full control points are used in the corners of the block.
This implies that there is not so much scope for reduction
of height control if the highest precision measurement is
required. This result was expected, knowing that reducing
plan control would still reduce the overall cost of a
photographic flight.
Control Point Standard Errors (mm)
3.0 2.1 1.9 1.6
3.1 3.0 2.5 1.6
4.0 3.2 2.9 1.8
7:5 N/A N/A
Table 4 - Combined Bundle Estimation With Different Control Configurations.
818
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996
ge
inr
The
poi
pre
1cr
usi
cor
poi
an
blo
Height SE (mm)
Fig
OVE
an:
cor
bei
dec
wh
Fic
