nually
ology
public
yrivate
ent in
iuse of
This
go had
ome a
nment,
DIGITAL PHOTOGRAMMETRIC STATIONS REVISITED
Armin Gruen
Institute of Geodesy and Photogrammetry
ETH Hoenggerberg
8093 Zürich, Switzerland
Commission II, WG IM
Key Words: Digital Stations, System Development, Design, Automation, Algorithms
Abstract: In the civilian market segment Digital Photogrammetric Stations are around for about one decade - as University-
based solutions first, later as commercial products. In some countries the digital systems have already outnumbered the
analytical plotters. This is an indication of the remarkable changes which digital technology has triggered even outside
research laboratories. This paper provides for a critical review of the state-of-the-art of current Digital Stations. It comments
on the achievements in system development and investigates to what extend the original expectations, demands and
predictions have been fulfilled.
1. INTRODUCTION
The first commercial fully digital photogrammetric
stereostation was presented as DSP1 by KERN & Co. AG
at the XVIth ISPRS Congress of the ISPRS in Kyoto, 1988.
It was wrapped into the PR-slogan ,We yodel digitally -
yodel with us“. The yodel with DSP1 did not last very long.
Problems with the provider of the integrated image
processing system and the mere fact that the system came
onto the market too early before the ,toe of growth curve“
(Helava, 1988) was reached, made it fail soon after. This
system had a number of military- and University-based
predecessors (Case, 1982, Albertz, Koenig, 1984, Gruen,
1986, 1989, Gugan, Dowman, 1986), so that as of today we
can speak of roughly a decade of concrete civilian
development of Digital Stations. This is the right time to
critically review the achievements in system development
and to investigate to which extend the original expectations,
demands and predictions have been fulfilled.
This review does not aim at completeness, it will rather shed
some spotlights onto individual system components. It will
express the views of a University researcher and teacher,
which are different from those of system manufacturers and
users. Since the author was one of the earlier proponents of
Such systems and has done with his group throughout the
years some research and development in this area, this
review will be personal and biased.
As a general observation, the relation between system
manufacturers and research groups has changed in recent
years substantially, as far as their attitude and approach to
innovative system and methodology development is
concerned. At the time of analogue photogrammetry and
the early years of analytical photogrammetry the system
manufacturers were the driving forces. Through their in-
house development of new instruments they provided the
subjects for the research groups in such a way that re-
searchers very often only could react and restricted them-
selves to testing and investigating the performance of these
instruments. With the advent of digital technology the
emphasis shifted in a twofold way. Firstly, creative new
ideas for algorithmic solutions in digital data processing
were required and secondly, through the use and
integration of inexpensive off-the-shelf hardware compo-
nents digital systems could be put together by virtually
anybody with sufficient knowledge in digital photogram-
metry. This led to the current situation where the software
solutions of manufacturers are trailing far behind the state-
of the-art of research. On the other side, many new system
vendors have entered the market, not all of them proving
that solid photogrammetric know-how in combination with
competence in image analysis can be transformed into
sophisticated and up-to-date solutions.
We have collected, in form of a table, information on 18
currently available commercial systems (excluding specific
systems for close-range applications). These are more
products than were available at any moment of time in the
history of analogue and analytical stereoplotters. This might
leave the impression that these are well developed systems
for a commercially attractive market. We cannot argue
about the latter, but the former definitely does not apply.
Development in close-range and aerial photogrammetry is
progressing along different tracks. Digital close-range
. photogrammetry has long gone fully digital and has majored
to an extend such that almost fully automated, robust multi-
image on-line solutions are being used in a variety of
practical application fields (industrial metrology, bio-
mechanics, animation, flow measurements, etc.). Digital
aerial photogrammetry on the other hand is still severely
struggling with automation, has analogue components in
the chain (aerial photographs, printed orthophotos and map
sheets), cannot quite separate itself from 2.5-D object
representations, as opposed to moving to a fully 3-D
approach, and is predominantly focussing on very
traditional two-image processing concepts. As an
explanation for this situation very often the higher
complexity and larger format of aerial photographs is
referred to. This is true to some extend. On the other hand
the mapping community, which is still the driving customer
force of digital systems, has never been overly innovative.
In this paper, we will exclude close-range systems, because
they constitute their own class of digital systems, with
application-specific requirements. We start with a brief
survey of commercial photogrammetric systems. Then we
will comment on some selected system components and
major functions, like image scanning, user interface, image
measurement/feature extraction, DTM generation, triangu-
lation, orthoimage production, monoplotting, and automa-
tion in general.
127
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B2. Vienna 1996
