- 509 -
A noteworthy aspect of present day use of remote sensing is the fact that
image analysis and information extraction are carried out by a variety of
techniques. These range from simple interpretation methods - similar to
those applied in many instances with aerial photography - to complex automatic,
semi-automatic and interactive digital processing, optionally coupled with
sophisticated information management subsystems. Choice of methods and
techniques depends upon the problems to be solved, scope of work and available
resources, equipment and expertise. The resulting flexibility contributes
significantly to the applicability of the data and the functioning of the
subsystems. Hence the importance to maintain in future such a diversity
in methods and flexibility in operations. However, it is not here the
place to discuss the merits and limitations of the various options (see
Refs. 2, 5, 8, 9 for details).
Nevertheless it might serve a useful purpose in the context of this paper
to refer to several significant developments with regard to the utilization
of imagery in a digital format. They concern the establishment of geo-based
information systems, for which satellite imagery with geometrical properties
approaching orthophotography can serve as a digitized spatial data base,
and the establishment of decision-models for the optimization of survey
procedures adapted to describing a statie situation in two or three dimensions,
as required (e.g. in mineral exploration), or to feature dynamic processes
(e.g. groundwater and plant-growth studies) by adding time as the fourth
dimension (see Refs. 2,.3, 9, 12, 13, 1h, 15, 16, 17). The specific inputs
which present and future remote sensing satellites could provide, are
mentioned in the litterature cited. In most cases these data have to be
combined with data from other sources, e.g. airphoto interpretation and
field observations.
It follows from the above that professionals in the various disciplines, and
experts working in inter-disciplinary or multidisciplinary fields, are in
a much better position to judge the relative value of various sensing,
image processing and information extraction techniques than 'remote sensers!
working more or less in isolation. They are also much better qualified to
integrate non-photographic remote sensing with the use of conventional aerial
photography in a problem-solving oriented approach.
This does not exclude the fact that there is an important role for a limited
number of remote sensing experts in translating end user requirements in
systems specifications and in carrying out the necessary R & D for systems
development and the design of the necessary range of hardware and software
for a dispersed and heterogenous user community.
What kind of remote sensing experts are required?
In the context of this paper remote sensing experts should be able either
to design or to evaluate sensing, image processing and information extraction
systems. They should therefore have an adequate background in physics,
micro-electronics and informatics. In addition they have to be creative
and capable to find novel solutions for new types of problems permitting
a full exploitation of the new dimensions which remote sensing is capable
of providing (Refs. 2, 5, 18).
This alone will not be sufficient. The experts should also be fully aware
of the complex interaction between technology and society, the consequences
of the lead-time between development of a technolegy and its acceptance
in society, and the non-linear, more or less cyclical transformation of
society as such,
Voute 5