of
ors?
1e,
am.,
)T,
Isat
ne,
lite
ne,
OT
ne
ne,
ST,
\R
ne,
lite
me
ne,
OT
me
ne,
lite
me
me
me
me,
OT
ime
ime,
allite ,
idar
ime,
POT
2. SURVEY OF COMMERCIAL SYSTEMS
Table 1 shows a summary of commercial photogrammetric
systems. It was prepared in essence by Dr. E.P. Baltsavias.
This account, although quite recent in nature (February
1996), does not claim completeness.
The following abbreviations are used in the Table to
describe the functionality of the systems:
AT; :5 «aerial
DAT ...digital aerial triangulation (semi-automated)
DC ...digital cartography
DM ...digital mapping. It includes feature acquisition and
updating, attributing and in some cases plotting
DTM ...acquisition (manual or automatic) of DTM
DTMm ...only manual acquisition of DTM raw data
GIS ...GIS functions. Usually for analysis, combination
and representation of raster data
GO ...geometric operations, coordinate measurement,
image warp/registration
GPS ...aerial triangulation with GPS camera stations
IP image processing. Most systems have rudimen-
tary image processing (LUTs, contrast enhance-
ment, histogram transformations, geometric
transformations), but some have a wider functionality
IS ...image sequence acquisition and analysis
MP ...monoplotting. Usually it refers to feature extraction
and vector update using an image (typically an
orthoimage) as a backdrop, without the use of the
underlying DTM for determining the height
OM ...creation of orthoimage maps. It involves combi-
nation of raster and vector data and plotting of the
orthoimage map
OP ...generation of orthoimages. In some cases it also
includes mosaicking
OR ..orientations (interior, exterior)
RS ...emote sensing functions (typically multispectral
classification)
SCA ...scanning of films by using a scanner of the same
vendor
TIN ...interpolation of a regular DTM. Derivation of other
products (contours, profiles, volumes) and their
visualisation
VI ... Visualisation. Typically it includes 3-D perspective
views of the terrain (overlay of orthoimages on
DTMs).
A closer look at these systems reveals that they are all more
or less consequently designed after the modular" concept,
recommended by Gruen, 1989 and realized by our group as
DIPS Il in 1986 and in the years to follow (Gruen, Beyer,
1991). The computer platforms are either Unix-based (SUN
or SGI workstations) or PC's. Special processor hardware
is used only in a few (expensive) systems. Stereoviewing is
predominantly through polarization and time multiplexing.
The majority of systems has been introduced fairly recently
(1992-1994). What this table does not show is a statistics on
out-of-production commercial systems. Since 1988 we
counted eleven such products, most of them equipped with
special processing hardware. This is a clear indication of
the fluctuations in this market. Currently, two major vendors
hold a share of 75 96 of the worldwide market. It will be
difficult for other competitors to alter this relation, unless
they can offer something more innovative, of higher
129
functionality and automation, easier to use and less
expensive.
3. COMPONENTS OF DIGITAL STATIONS
3.1 Image Scanning
Image scanners are offered nowadays at very different
levels of functionality, performance and price. Baltsavias,
Bill, 1994 give a good account of scanners suitable for
photogrammetric applications. Scanners are currently
probably the only area where Digital Stations-related
scientific investigations are being conducted at a larger
scale and with international participation, including equip-
ment manufacturers (OEEPE/ISPRS Working Group on
Scanner Test). It has been shown by Baltsavias, Waegli,
1996 that medium and low cost desktop scanners, if
appropriately calibrated, can produce results almost as
good in terms of geometric quality as the highend
Ë’photogrammetric“ scanners, with partially superior radio-
metric quality. Another distinctive advantage of desktop
scanners is the ease of use and the high scanning speed.
This has to be weighted against restrictions in geometrical
resolution and format. Currently flatbed desktop scanners
cannot produce data with pixel sizes smaller than 21
microns. If, however, the development of general purpose
scanners will turn out to be as dramatic as in the field of
printers we may very soon witness a substantial improve-
ment in resolution and a further drop in price with the result
that specific ,photogrammetric“ scanners are not of interest
any more. Scanners also should offer some local proces-
sing capabilities. Functions for image enhancement,
editing, browsing, and image resampling are already
available in form of public domain raster processing
software. Also, interior orientation should already be
established and performed automatically at the scanner
level, before the data is transferred to the Digital Station
(compare e.g. Leica/Helava DSW 200, Zeiss/Intergraph
PS1). Finally, some level of on-line scanning control and
local intelligence could be imagined, which allows to scan
local image regions with varying geometrical resolution,
either according to preset locations (e.g. for locally high
resolution signalized object point regions), or even by on-
line analysing the signal content of image regions and
adaptively changing the resolution. These image data sets
would of course also require particular storage and data
handling techniques at the Digital Station. Another option
would be to scan at very high resolution and to apply an
adaptive compression technique which maintains the high
image frequencies (e.g. signalized points) locally.
3.2 User Interface
In a largely automated system the user interface should not
play a very significant role. Helava, 1988 remarks that
Digital image photogrammetric systems should be
designed to be automation friendly“. The current status of
the user interfaces of the commercial systems is actually a
strong indicator of the low level of automation. Buttons,
menus, scroll bars and images are all over the place in most
systems. It needs a real knowledgeable and experienced
operator to navigate without failure and frustration through
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B2. Vienna 1996
