positioning data from CPNS /12/ and can be considered a gene-
ral feature of combined adjustment.
In the following some simulation results are discussed concer-
ning joint adjustment of navigation data and photogrammetric
blocks. They are intended as a first overall look at the mat-
ter, in order to see what effects can be expected from diffe-
rent types (position, attitude) and different accuracy levels
of camera orientation data. The simulation is very much sim-
plified, It is not realistic with regard to some items,
omitting, for the time being, all detailed considerations of
drift, systematic errors, and correlation of observations.
We know from experience that more detailed investigations -
which will certainly have to follow - will not change the
overall relationships very much which are obtained from sim-
plified error models (although second or higher order drift
properties would have consequences with regard to control
requirements).
Reference is made here to the simulations by P. Friess the
results of which are presented during this symposium in the
session of Working Group III/1, /13/. Most of the simulations
refer to photo scale 1:60 000, block size 10x41 photographs,
idealized geometry, precision of image coordinates oy = 15 um,
bundle blockadjustment, including external orientation para-
meters as additional observations. The actual assumptions,
however, are not limiting. What really counts are the weight-
relations involved. Therefore, the results can easily be
transferred to other cases. The further details are descri-
bed in /13/. Here the major results are summarized in fig. 1
and will be briefly commented:
4.2 The overall result is that external orientation data
stabilize a block very much. They are highly efficient in re-
ducing control to the point that ground control is not at all
necessary anymore for accuracy reasons. This is true even
for navigation data of relatively poor precision. Control is
reduced to the function of solving for datum and configura-
tion defects, i.e. of establishing the coordinate reference
system and, if necessary, of determining drift parameters and
systematic errors. Apart from that, blocks could be adjusted
without any control, referring to the reference system of the
auxiliary data.In fig.1 all results refer to 4 xyzcontrol points
per block, which may be considered as standard for auxiliary
data.
In /13/ it is shown that additional ground control do not
improve the block-accuracy very much, especially not when
auxiliary data are in the high accuracy range.
Remark: The more accurate auxiliary data will be the more the
problem of interior orientation of photographs will become
predominant. This problem is excluded here from further con-
siderations. There will be several ways of handling the prob-
lem, for instance with additional parameters. This aspect
will require further investigations.
