Maps Attr ibutes
| 1
Graphics Input [Attribute Management Attribute Output
Graphics Pre-process
1 |
Graphics Operation||Image Operation
Digital Terrain Model
Geo-model Analysis
Graphics and image output
Fig. 4 The structure of MCGIS.
MOGIS uses Geo-Code Model (GCM) to describe
geographical attributes.GCM reasonably grades
geographical factors and uses symbol strings,
such as digital strings and literal strings,
to express geographical attributes. Every bit
in those strings represent different
attribute, with its value representing the
attribute value.
Vector and raster (grid) are two basic
graphics data formats used in MCGIS. Vector
data is mainly used to input graphics and
calculate geometric data (area, length etc.),
and raster data or images is mainly used to
operate graphics and images.MCGIS defines two
basic image operation, that is, image overlay
and image extract, both of which are based on
the GCM-expressed geographical attributes.
Image overlay can be expressed as
11+12+13+...+In —> I
where
(1)-11, 12, ..... ,In are as independent
geographical factor images (raster maps);
(2) I as an integrative geographical factor
image,which is composed of I1, I2,...,In.
Image extract is the reverse calculation of
image overlay,which can be expressed as
I -»"Ti1-I2TI2t0... 1n
that is,getting some dependent or independent
geographical factors from an integrative
factor.
Generating Digital Terrain Model (DTM) is
another important function of MOGIS. DTM can
be expressed as
Ti=Fi(x,y)
where
(1) Ti as topographical factors, such as
elevation,aspect,slope etc;
(2) Fi as the function of topographical
factors to geographical coordinates (x,y).
MCGIS also uses a series of mathematic
methods to solve geographical problems with
271
known models.Those mathematic models includes
multiply regression analyses model, tendency
analyses model, cluster classification model
and principal component analyses model and so
on.
2.2 GKB — geographical knowledge acquisition
and management system of MCGES
As a branch of AI applications, expert system
uses knowlege as logical kernel. Geographical
knowledge is the logical association of
geographical data, such as, geographical
planning, experience and phenomena.Therefore,
geographical knowledge must be acquired from
geographer. The level of acquisition and
management of geographical knowledge is an
important factor to show the level of GES.
Knowledge acquisition includes knowledge
representation and knowledge refinement, and
the former is the most important one.
Geographical knowledge representation may be
defined as expressing geographical knowledge,
such as planning, experience and phenomena,
with abstract logical languages that can
be understood by computer. This kind of
expression can be inputted to computer and
stored in knowledge base.
After analysising geographical research
contents, we devide geographical knowledge
into following three levels:
(1) the basic level,which describe and grade
the geographical independent factors, is based
on GCM;
(2) the medium level, which classifies
integrative geographical factors, is a group
of classification rules;
(3) the advanced level is a series of
integrative geographical planning rules.
A kind of list language is designed to input
geographical knowledge in MOGES,and there are
several functions in the knowledge base
management module or sub-system (see Fig.5).
2.3 GIE — reason controlling module of MCGES
Inference engine controls the inference
operation of the knowledge in the knowledge