Improvement in planimetric accuracies was not 
noticed in these experiments, probably because the 
large-scale block used has a high planimetric 
accuracy (vithin 10 cm) when it was conventionally 
adjusted. Improvement in planimetry was observed 
in the experiments with generated data and in 
particular in medium size blocks (10 strips, 40 
models/strip). The use of only four ground control 
points, without additional height points, plus GPS 
data does not yield satisfactory results , while 
the addition of one XYZ point in the block centre 
gives much better results. It should be noted, 
however, that these control configurations may 
cause numerical instabilities in practical 
applications where the blocks are not symmetric or 
as well prepared as simulated or test field 
blocks. 
From the different control configurations used in 
the combined adjustments, the most effective with 
respect to block precision and yet with 
substantially reduced ground control points is 
four XYZ points at the block corners and a chain 
of height controls at both the beginning and end 
of the block. In aerial triangulation using GPS 
data, there is no need for ground control points 
in the inner area of the block. 
It is clear from the presented results that, GPS 
modelling with constant, linear and quadratic 
terms worsens the accuracy and in particular the 
height accuracy. It seems that the block 
deformation in heights adapts itself to the GPS 
modelling rather than being controlled. On the 
other hand, GPS modelling with constant and linear 
terms gives a strong support for controlling the 
height block deformation. 
The statistical test applied for detecting if 
parameter values are significantly different from 
zero is shown to be very effective, and provides 
the means of avoiding the detrimental effect of 
insignificant parameters. The test to determine if 
parameter groups are significantly different from 
each other should also be incorporated in combined 
adjustments software, thus avoiding the 
introduction of an excessive number of parameters 
in the system. 
It is evident that the high accuracy and 
functionality of the kinematic GPS will require 
the photogrammetric community to reconsider the 
planning of photogrammetric projects. 
REFERENCES 
1 Ackermann, F. 1988. Combined adjustment of 
airborne navigation data and photogrammetric 
blocks. In: Proc Comm III ISPRS, Kyoto. 
2 Andersen, 0. 1989. Experience with kinematic 
GPS during aerial photography in Norway. In: 
Proc 42nd Photogrammetric week Stuttgart 
University. 
3 Bouloucos, T. 1986. Multidimensional tests for 
model errors and their reliability measures. 
ITC Journal 1986-3 
4 Colomina I, 1989. Combined adjustment of 
photogrammetric and GPS data. In: Proc 42nd 
Photogrammetric Week, Stuttgart University. 
5 El-himdy Abdelilah 1989. Block adjustment with 
independent models using global positioning 
system data. MSc thesis, ITC Enschede. 
6 Friess, P. 1988. Empirical accuracy of 
positions computed from airborne GPS data. In 
Proc Comm III ISPRS, Kyoto. 
508 
7 
10 
11 
Friess, P. 1991. Aerotriangulation with GPS 
methods, experience, expectations. In: Proc 
43rd Photogrammetric Week, Stuttgart 
University. 
Kleusberg A. 1991. Principles and performance 
of kinematic GPS positioning. In: Proc 43rd 
Photogrammetric Week, Stuttgart University. 
Li Deren, Shan Jie 1988. Quality analysis of 
bundle block adjustment with navigation data. 
In: Proc Comm III ISPRS,Kyoto. 
Vegt H.J.W van der. 1989 GPS test flight 
Flevoland. In: Proc 42nd Photogrammetric Week, 
Stuttgart University. 
Witmer R. 1988. Toepassing van GPS in de 
Fotogrammetrie. Afstudeerscriptie, Technische 
Universiteit Delft, The Netherlands. 
proc 
ima 
proc 
hist 
paci 
only 
subı 
be 
pare 
alth« 
man 
capt 
softv 
prog 
help 
pack 
prod 
expe 
to b 
techi 
such 
quali 
is th« 
is de 
class 
evalı 
and i 
error