de
ne
| 2
V
je
at
le
nt
le
e
id
Le
id
je
ar
f
features is essential. It is not sufficient just to detect features/changes
the exact positions must also be known. AIMS has the capability to merge
digital radar with the Defense Mapping Agency's(DMA) existing point
positioning data base(PPDB) The AIMS user will have the ability to
transfer points of interest from the digital radar to the PPDB and precise
latitude, longitude and elevation data from the PPDB to the radar scene to
satisfy extremely accurate positioning requirements. This information is
of great to the military e vertical obstuctions, radar emitters, bridge
crossings, and overall target positioning).
ARTIFICIAL INTELLIGENCE IMPLICATIONS
Aims will be implemented at Autometric using a conventional
information system design and development techniques and eventually a
more robust set of artificial intelligence (Al) techniques. he
subdisiplines of Al the will be used include 1) imagery understanding, and
2) expert systems.
Imagery Understanding
Imagery understanding algorithms will be applied to the features that
have been detected in the digital radar imagery. The feature extraction
algorithms will seperate features from their backgrounds achievind a
'stable structure' (Hsu, 1985). Imagery understanding software will then
be applied to these detected features. Inferences are made from
recognizable shapes and patterns.
Expert Systems
Expert system techniques will be used to support applications that
require interpretive, diagnostic,predictive,prescriptive or planning
assistance. Expert systems are currently being developed at Autometric
using cartographic feature rules for the automated analysis of radar
imagery. The basis of the expert system is a knowlodge base containing
inference rules using radar feature shape/pattern descriptors together
with digital MC&G feature information. As new knowlodge is aquired
and/or new inference rules are developed they will be added to the
knowlodge base and inference rules set without having to change the basic
expert system architecture.
CONCLUSION
It appears that the necessary technological hardware and software
environmnent exist to make AIMS fully operational as an exploitation
system of SAR as well as other imaging sensors. AIMS has the unique
capability of being able to handle radar imagery in both analog and digital
formats. The user is able to take advantage of the more beneficial
exploitation potential of either format by utilizing AIMS modular software
and hardware configurations.
Although digital image processing techniques are rapidly advancin
using artificial intelligence methodologies, digital imagery still can no
stand alone as a completely exploitable mae format. Perhaps the time
will come that all imagery will be in a digita format, but until that time,
«AM will serve as an ideal transitional system between the two image
ormats.
369