3ssungsamt 
ing a Zeiss 
15 um pixel 
pt for some 
or, this very 
monstrating 
ias decided 
it seems to 
data set of 
y 2x2 pixel 
S 
o perform 
data sets. 
eties of the 
jerning the 
adjustment 
usted using 
) control as 
sing original 
mparison to 
PANTS 
pelow, have 
Norway 
Switzerland 
Switzerland 
Finland 
Finland 
France 
Sweden 
Sweden 
Germany 
Austria 
Canada 
Germany 
Norway 
djustments. 
ch can be 
asurements 
as on an 
eoscopic or 
onventional 
n treatment 
are mainly 
this group 
ome form in 
. The most 
gous points 
ask is the 
measurement of fiducial marks and signals, mainly using 
template matching. Essential in this group is that the 
approximate values of points must be known in advance 
or gained using manual pointing. 
3. Automated methods ( AM ). The basic difference, as 
compared to semiautomatic methods, is the automatic 
selection of tie points and acquisition of approximate 
image coordinates. In its most advanced form, the 
method is practically fully automatic. In the current status, 
at least the measurement of signalised points requires 
human interference. 
In the following, the naming of the cases has been 
grouped according to the methods or measurements. 
Within the ranges, the cases usually differ in pixel size 
and type of block adjustment. 
École Polytechnique Fédérale de Lausanne ( EPFL ) 
The measurements have been made using Intergraph 
ImageStation and Leica DPWsystems. Both are 
commercially available semi automatic systems. The 
measuring resolution of DPW system is one pixel. The 
adjustments have been made in all cases with additional 
parameters. 
The BC1 and DSW100 ( with DCCS software ) 
measurements have been made for comparison. 
EPFL1-2 Wild BC1 analytical plotter measurements 
using analogue imagery. The cases have 
often been referred to with the name BC 1. 
(VM) 
EPFL3-4  Helava DSW 100 measurements. ( SM) 
EPFL5-6 Leica DPW measurement. ( SM ) 
EPFL7-8  Intergraph ImageStation measurements. 
(SM) 
EPFL 9- 12 Adjustments with natural tie points. 
Eidgenössische Technische Hochschule Zürich 
(ETH) 
The measuring system was the selfdeveloped DIPS II. 
The semiautomated system includes also the automatic 
measurement of fiducial marks and signalised points 
using automatic template matching. (SM) 
Finnish Geodetic Institute ( FGI ) 
The used system was a  selfdeveloped digital 
photogrammetric workstation, where automation was 
used in image matching. The matching is based on a 
cross-correlation method. The block was measured using 
different approaches: 
1. Completely visual or semiautomatic measurements. 
2. Different densities of tie points ( 17 - 34 points/image ). 
3. Different number of visual observations per signalised 
ground point. 
4. Use of natural or signalised points as tie points 
FGI 1-8 Visual observations. ( VM ) 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996 
FGI 9-16 Points were measured visually on the first 
images. On the following, correlated with 
the first observations. ( SM ) 
FGI 17-24 Like FGI 9 - 16, but the tie points were 
natural details ( no signals ). ( SM ) 
FGI 25-26 Like FGI 9 - 16, but all the points have 
been measured visually and then 
correlated to corresponding observations 
on other images. ( SM) 
FGI 27-34 Like FGI 9 - 16, but the number of tie points 
( natural ) has been increased from 27 to 
34 per image. ( SM) 
FGI 35-36 Like FGI 25 - 26, but using tie points like in 
FIG 27 - 34. (SM) 
Helsinki University of Technology ( HUT ) 
The experimental system was developed in the institute. 
For  monoscopic visual measurements, variable 
measuring marks were available. The block adjustments 
were performed without additional parameters. ( SM ) 
Institut Géographique National ( IGN ) 
The 15 um data was measured using a selfdeveloped 
system on a Vaxstation ( IGN 1 - 4) ( SM ). The 30 um 
data was measured on a Matra Traster T10 ( IGN 5 - 8 ). 
(VM) 
Kungliga Tekniska Hégskolan ( KTH ) 
Measurements were made completely visually using 
ERDAS software ( KTH 1 - 4) (VM ) and 
semiautomatically using a system developed in the 
institute ( KTH 5 - 9) (SM). The semiautomatic system 
includes template matching on fiducial marks and signals. 
Agricultural University of Norway ( NLH ) 
The measurements were made visually using ERDAS 
OrthoMAX block triangulation software. The 
measurements were made without stereoscopic viewing 
and area based image matching, even if they were 
included in the system. ( VM ) 
National Land Survey of Sweden ( NLSS ) 
Measurements were made visually on Teragon system. 
(VM) 
: University of Trondheim ( NTH ) 
The measuring system was ERDAS with the remote 
sensing module. (VM) 
Technical University Munich ( TUM ) 
The measurements were made using an experimental 
system. The observations were made visually with two 
measuring resolutions, 1 pixel ( TUM 1 - 4 ) and 1/3 pixel 
( TUM 5 - 8 ). One control point in dense configuration 
  
  
ER REE