International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
Additionally a thorough accuracy control for each block was
carried out, which was done by using check points. The
reference coordinates of these points were either terrestrial
surveyed or taken from numerical photogrammetric mapping
projects 1:1.000. These data were part of the data base projects
for community-government boards (see 1.).
4.2 AAT results using inBLOCK
Table 4 shows the AAT results of the complete adjusted block.
Strips / images 144 / 2447
Full control points 259
Height control points 152
Sigma 0 — image
Sigma 0 - full control points
Sigma 0 - height control points
RMS image points
RMS control points
0.41 / 0.40 [um ]
0.136 / 0.138 / 0.156 [m]
0.182 [m]
2.895 / 3.345 [um]
0.156/ 0.129/ 0.180 [m]
Table 4. Complete block, results of block adjustment with
: +
inBLOCK
VE : BR ubt e eS «mx
d cus Hakan, pea pese qs tir
ce A
sie au —
mhi se zr zx wl
| = i op a =
a9 ; * ©
| = x ü a 2
*
ÿ n.
i seh : - x $8 3
ai EE a 2
| Dés ER n ac + a
i Y e a be cR EP * - à =
Pu E al E + &
ak * ES = VM 5 "n
| o * RE a Am us m ¥ -
i a P$: s» > E
i + = » & 2 * v =
i § € E 4 = H
| * » Beh y
| : 5 = * ud = "à I ir
| # we se + =
| »* RE 4 T^
NE were.
i ww
i - ec OX E. +
m a som, A
n > ma ®
i * % en
i FF e.
at “
T T s
| ; a x x
: a 1 -
| pe a
| -5% X
i z z
| +. 8
| nix
i
A Se in Yan
N io jme: emm] visae e
Figure 2. Overview of complete block with control points, new
points and strips
For analysing the block lots of tools are available. One of those
is the so-called blunder browser. The blunder browser is used to
display information about possible blunders or eliminated
blunders. The dialog consists of three parts: the numerical
panel, the graphics panel and the control panel.
The numerical display shows the critical points (possible or
eliminated blunders) with all observations. If there are blunders
detected for position or attitude observations for projection
centres, these are displayed as well. It shows the test values
(normalised residuals) and estimated errors for the observations.
The graphics display shows the estimated errors for the selected
blunder in the numerical panel.
So the block can be checked very detailed. The weak areas of
the block are found, checked and improved easily.
In inBLOCK all residuals and/or estimated errors of all points
(control points, image points, GPS- or IMU-observations) can
be displayed. For the current project no problems in the image
636
measurements were reported. The problems which had to be
solved appeared in the control points. The GPS-data did not
improve the result because the accuracy of 1m was too weak.
So the control points were of particular importance. We
checked the strength of the control points. Here the estimated
errors in comparison to the residuals provide an informative
basis. The estimated error shows the size of error the point
could have without detecting it as *blunder'.
Furthermore we checked the standard errors of the block. Here
the error ellipses show the maximum error (calculated from the
block geometry, number of measurements and the actual
residual). The block had a homogeneous result, the error
ellipses at the border are larger (2-fold points), the center of
block showed only small error ellipses.
Also the internal reliability of the block was checked, which
represents the size of a minimum detectable gross error for an
individual observation. It can also be seen as a measure of the
controllability of the observation.
External reliability is the influence of non-detectable gross
errors onto the estimated parameters. It can also be seen as a
sensitivity measure or robustness of the network against
distorting effects in the system. Gross errors smaller than the
threshold values (internal reliability figures) may stay
undetected and contaminate the result. It is interesting to
analyze the maximum influence of undetectable errors onto the
estimates.
Inspecting the larger ellipses of the block showed that they are
mainly caused by 2-fold points. These weaker points are
checked by neighbouring points that had better analysis values.
All information’s such as coordinates, standard deviations,
correlation matrix and outer reliability of the point can be
displayed.
5. CONCLUSIONS
The adjustment results of MATCH-AT and inBLOCK gave
similar results. The block was checked optimal by the analysing
tools of both programs. To optimize the processing it would be
preferable to use GPS-data in a higher accuracy (e.g. 0.10 —
0.15 m).
This report shows that using
- an appropriate software
- a correct pattern of control points within the block
- andthe necessary Know how for carrying out such
AAT projects
a high accuracy can be achieved even under difficult
circumstances as given in the presented project.
Closing this paper it shall be mentioned, that the project of part
1 is finished meanwhile and that Provincia Autonoma di
Bolzano — Alto Adige ordered the bidding group GEOSIGMA
srl, GEOMATICA srl and Ingenieurgemeinschaft
Vermessung AVT-ZT GesmbH to start with part 2 of this project
(see Figure 1.).
References from Books:
Kraus, K., 2003. Photogrammetrie, Band 1, 7. Auflage. Walter
de Gruyter, Berlin, New York, pp. 285-306.
Inte
Sail
Pho
Wic
Ref
Hei
extr
Pub
