5. In general, when choosing 0. = 0 = 10u as a priori standard
deviation for all tie-points and’c,=2 in the linear choice function
as shown in formula 3.1 an error of 30 cm in one of the control
coordinates can be traced «easily.
However, when decreasing O and O_ to 54 the impact on the boundary
values of the control points is significant (see table 6).
Comparing the eigen value results of experiments 5 and 6 the same
effect is clearly shown.
The overall conclusion may be that if the control points are spread
out along the periphery of the block the impact of those points on
the final result of the block adjustment in terms of reliability and
precision is significant. However, it is also clear that if it is
possible to decrease the a priori standard deviation of the model co-
ordinates, the final result becomes rapidly better.
REFERENCES
/1/ Baarda, W.: A testing procedure for use in geodetic networks.
. Netherlands geodetic commission, Delft. 1968.
/2/] Baarda, W.: S-transformations and criterion matrices,
Netherlands geodetic commission, Delft. 1973.
/3/ De Heus, H.M.: Precision in case of combined point determination.
Paper presented at ISPRS commission III symposium
Rovaniemi 1986.
/4/ Van Mierlo, J.: S-transformations. internal report Delft University
of Technology, department of geodesy.
/5/ Molenaar, M.:. The connection of aerotriangulation blocks re-
. considered. Paper presented at ISPRS congress,
Rio de Janeiro. 1984.
/6/ Molenaar, M., Karandis, D., Bouloucos, T.: A substitute matrix for
photogrammetrically determined point field. Paper
presented at ISPRS congress, Rio de Janeiro.
/7/ Molenaar, R.C.: The inertial reliability in a 3-Dimensional strip
adjustment with independent models. Paper
. presented at ISPRS congress Hamburg. 1980.
- 68 -