efining
arrying
of the
1g true
ounded
)f data
ms to
rithms
nes or
an be
mated
| plane
rfaces
es.
efined
of the
ays as
ead of
can be
object
SW, Or
ontour
rge to
' such
d and
Kager,
sts as
could
ng the
ng to
to the
'ation-
screen
frame
ill the
lasses
Upon
those
d and
phics
class
it can
] case
forms
file of
| type
ta, as
class
erent
1 and
DTM Algebra applies as operands different DTMs of the
same area carrying different epochs of the same
information (e.g. elevation), or different types of
information (e.g. slope or soil quality models). There
may be used very complex expressions. This is one of
the functions of the current SCOP.INTERSECT module
(Sigle, 1991), parts of which should be adapted as a
method of this class. C/assification is to incorporate the
second important function of SCOP.INTERSECT
allowing to classify contents of a DTM according to a
network of polygons of any complexity (e.g., to classify
a slope model according to cadastral boundaries). D7M
Exploration is for compiling (deriving) different products
o S DM. by methods such as reported in (Rieger,
It is of mandatory importance to represent results of
different methods not just as graphics or listings but in
forms capable of further processing - such as, again,
DTM structures with the derived quantities as z values.
Slope models are a good example of this, but even
visibility of the surface from a specified point of space
(a by-product of the hidden line algorithm of perspective
views) can be represented as a DTM structure. In this
case, the elevation difference (negative, O, or positive)
should be represented as z, to be added to the elevation
of the DTM location so to become visible or invisible,
respectively (Hochstoeger, 1991).
SYSTEM INTEGRATION
The classes as described will be integrated into two
different versions of the DTM system:
- a full-featured version with graphical user interface
(GUI), command language, interactive graphics,
etc., available for some of the most important
platforms (such as PC workstations under OS/2
2.* and under WINDOWS NT, some major UNIX
workstations, and maybe DEC VMS), and
- a major machine-independent subset of the DTM
system, mostly for mainframes, with just the
command language as user interface.
Both versions will run in three modes of operation:
- interactive mode,
- batch mode controlled by command procedures,
and
- driver or slave mode, to service requests of a host
system, without being seen by the user in any
ways (him seeing just the user interface of the
host system), and sending all output to that
system.
Integrating the DTM system into a greater application
software environment concerning both input and output
is of crucial importance. A new edition of the
Topographic Information and Archiving Software
(TOPIAS) based on TOPDB is to be integrated into the
DTM package so to play the role of some foreign
secretary. Messages in the syntax of standard SQL, or
better still, of TOPSQL should enable communication in
both directions. Servicing (TOP)SQL requests by the
DTM system yields a good example of "lazy
processing": requesting (SELECTing) an elevation at a
location (x,y) will result in a message to the class DTM.
This will check whether there is an interpolated surface
ready at the location; if not, a message will check with
the class of input data whether interpolation is possible,
and so on. After some "small talk" among the classes
exchanging messages and data, the result will be
deduced and sent back (as INSERT statements) via
TOPIAS to the requesting system (which could be a
user-written program, as well).
The driver or slave mode of operation provides fine
means of integration with geographic information
systems (GIS), with systems to serve analytical plotters
(e.g. to support data acquisition: progressive sampling,
editing, image overlay of different DTM products), or
with interactive graphics systems.
And furthermore, there remain such means of
integration as DXF and other file format standards.
Compatibility with earlier versions of SCOP must be
ensured by supporting not only the corresponding 1/0
formats but also both databases (DAF and RDH).
SUMMARY
An architecture for application software is described, as
proposed for the new edition of the DTM package
SCOP. It consists of fairly independent modules
("autonomous classes’), integrated within an object-
oriented frame. Autonomous classes can be
implemented, in addition to object oriented
programming, in classical procedural ways. This
compromise has major merits at the current state of the
software engineering practice: it provides fair portability,
allows for integrating modules of earlier programs, from
the user's perspective it carries clear signs of object
orientation resulting in user friendliness, and the system
is easily extendable by further autonomous classes.
Some selected classes of the DTM system are
described. Further important application classes are
going to be created as important by-products of
dissertations, the commercial benefits providing the
student with means for his studies, and the application
helping to spread new technology.
Acknowledgment
This study has been financed by Fonds to Promote
Scientific Research of the Austrian government, Project
Nr P7385-GEO.
References
Assmus, E., Kraus, K., 1974. Die Interpolation nach
kleinsten Quadraten; Praediktionswerte Simulierter
Beispiele und ihre Genauigkeiten. Deutsche
Geodaetische Komission, Reihe A, H.76.
Ecker, R., 1991. Rastergraphische Visualisierungen
mittels digitaler Gelaendemodelle. Geowiss. Mitteilungen
der Studienrichtung Vermessungswesen der TU Wien,
Heft 38.
Hochstoeger, F. 1989. Ein Beitrag zur Anwendung und
Visualisierung digitaler Gelaendemodelle. Geowiss.
Mitteilungen der Studienrichtung Vermessungswesen
der TU Wien, Heft 34.
Hochstoeger F., Ecker R., 1990. Application and
Visualization Techniques for Digital Terrain Models.
International Archives of Photogrammetry and Remote
Sensing, Komm. IV, Vol. 28, Part 4, Tsukuba, Japan.
Kager H., Loidolt P., 1989. Photomontagen im
Hochbau. Vermessung, Photogrammetrie, Kulturtechnik,
Nr.3.
Kalmar, J., 1991. Diverse Interpolationsverfahren.
Institut fuer Photogrammetrie und Fernerkundung, TU
Wien. /ntern.
Koestli, A., 1988. Manual for BC FORTRAN. BC
Software, Bad Canstatt bei Stuttgart.
967