$$ o
THE POTENTIAL OF HYPERMEDIA TO PHOTOINTERPRETATION
EDUCATION AND TRAINING
D.P. Argialas O.W. Mintzer
Remote Sensing Laboratory U.S. Army TEC.
Dept. of Surveying Engineering ATTN: CEETL-RI-T
National Technical University 2592 Leaf Road
9 Heroon Polytechniou Str. Fort Belvoir, VA 22060-5546
Zografos 15773, Greece USA
Commission VI
PURPOSE:
Hypermedia is an authoring environment, an idea organizer tool or a conceptual indexing tool. A
hypermedia information base consists of interlinked pieces of text, graphics, full motion video,
animation and sound. The authors examine how educators teaching photointerpretation could use
this technology for more effective presentation and teaching. A photointerpretation hypermedia
system can help put information related to objects, attributes, their clues and recognition methods
in obvious patterns so students could interactively and non-sequentially locate, browse, and
understand the photointerpretation concepts and processes. The marriage of multimedia and expert
systems is also inevitable. Hypermedia systems can help to build less ambiguous decision
support systems and knowledge-bases and thus make expert interpretation systems more intuitive.
KEY WORDS:
hypercard, teaching
INTRODUCTION
Photointerpretation is still an art, using
science for only its most basic foundations.
Practice relies on the case-study method and
intuitive judgement more heavily than is
commonly believed. Photointerpretation
relies on an imprecise combination of
information gathered in different ways and
processed in a largely intuitive and recursive
fashion. Overall impressions are gathered
and interpretations are made, but the
photointerpretation process remains obscure
(Argialas et al, 1999). "There are few
definitions and descriptions of
photointerpretation objects given in the
literature. And, there are even fewer
explanations of the procedural or inferential
framework of the interpretation process
(Argialas and Narasimhan, 1988a;
Narasimhan and Argialas, 1989). This may
be so because of the difficulty of the problem
itself, the lack of theory and technology to
assist in the symbolic representation of the
complex inferential process (Argialas and
Narasimhan 1988b), and in the extraordinary
emphasis that has been placed on the digital
image processing paradigm. The last reason
has contributed to our sidetracking from
forming a workable photointerpretation
paradigm (Ryerson, 1989). For the last two
decades little attention has been paid to
rigorous, visually-based image interpretation.
This severely reduces our ability to
effectively interpret and use the TM, SPOT,
and other sensor images.
There is, therefore, a need to study
methodically the process of building
photointerpretation keys and to develop a
systematic framework for the recognition of
objects from aerial images using information
specific to how objects appear on a given
image as well as information available in the
375
Photointerpretation, hypermedia, expert systems, knowledge representation
mental models of expert photointerpreters.
The proper application of such keys will
allow a novice interpreter to tap onto the
knowledge base of the individual who
offered her/his knowledge for building the
key. To build such keys it is necessary to
have a thorough understanding of the domain
specific knowledge and of the new theories
and technologies in representing, managing,
retrieving and formalizing knowledge
(Nielsen, 1990; Shneiderman and Kearsley,
1989; Parsaye et al, 1989;Bielawski and
Lewand, 1990). Advances in hardware and
software has contributed to a new set of
techniques for computer aided instruction and
presentation systems of unparalleled impact.
Hypermedia and knowledge-based expert
systems can offer us the base on which to
develop a workable visual image
interpretation paradigm similar to the keys of
the past. Argialas and Harlow (1990) have
examined the potential of expert systems for
image interpretation. In this paper the authors
examine the potential of hypermedia for
photointerpretation and present the rationale
underlying the structure and design of such
systems.
THE HYPERMEDIA CONCEPT
Hypertext was born of the need to devise text
storage and retrieval systems capable of
managing efficient large masses of
information. All traditional text, whether in
printed form or computer files, is sequential,
that is there is a single linear sequence
defining the order in which the text is read.
The idea behind hypertext is to read text non-
sequentially. Hypertext has departed from
the sequential, linear storage and retrieval of
text to a random-access, nonlinear method
with many options and alternatives (nielsen,
1990). Hypertext can present several options
to its readers, and each one determines which