Conceptual Model
On this page we present
the data model of SPRING, which describes how the geographic reality will be
represented in the system. The greatest contributions of the model are:
-
integrate Remote Sensing imagery and Digital Terrain Models with
thematic maps, cadastral maps, and networks.
-
Define a mapping between geographic objects and their
localization, allowing for more than one geographic information to be
associated to one same entity in the real world.
-
provide a high level interface with semantic content.
-
allow the coexistence of vector, raster and grid data in the same
system.
A data model is set of
conceptual tools that is used to structure data in a computational system. A
fundamental aspect in the design of a GIS like SPRING, the model describes
how the geographic reality will be represented inside the computer. No other
decision will impose tighter limits on the comprehensiveness and future
growth of the system like the choice of the data model.
The topics discussed now are:
See also:
How
to define a database in SPRING
Notions
of Geoprocessing
About
SPRING
How
to define a data model in SPRING
Object Orientation in
Geoprocessing
The term object
orientation denotes a work paradigm that has been widely used in the project
and implementation of computer systems. In this section we offer a short
introduction to the theme, providing some theoretical support to the
understanding of the SPRING data model.
The general idea of the object-oriented approach of a problem is to apply the
classification techniques. The two fundamental concepts of object-orientation
are class and object. An object is an entity that has a description
(attributes) and an identity. A class encompasses objects that share some
common properties. For example, in Biology the class of Mammals
encompasses all the animals with the properties of having a hot blood,
and that are breast fed by their mothers. In this case one can say about
"monkey Joe" that "monkey Joe is a mammal".
For a more thorough analysis it is useful to recognize sub-classes, derived
from a basic class that allow for a more detailed analysis. This mechanism is
named specialization. Thus we can say that the "Primate" class is
an specialization of the class "Mammals". Such process could go on,
and we could further define a "Human" class as a specialization of
the class "Primate".
In this process of specialization the derived classes inherit the properties
of their basic classes, adding new attributes that are specific of these new
classes. As a consequence, the statement "every human is a mammal, but
not all mammals are human" holds.
Another fundamental mechanism in the theory of object-orientation is
aggregation (or composition). A composite object if formed by the
grouping of objects of different types. Take the case of a computer. We can
think of a computer as being an object composed of a CPU, memory, hard disk,
keyboard, monitor, and mouse.
Object oriented modeling applies naturally to Geoprocessing. Each of the
spatial data types will be described a class that could obey to some sort of
hierarchy, where the derived sub-classes would inherit the behavior of the
more general classes.
In Geoprocessing the idea of specialization (also called "is-a") is
normally used to define sub-classes of geographic entities. For example, in a
cadastral map the class of objects identified by hospital could be further
specialized in public hospital and private hospital. The attributes of the
class hospital are inherited by sub-classes public hospital and private
hospital, that can have specific attributes themselves.

Figure
exemplifying specialization.
The relationship of
aggregation (also called a "part-of" relationship) allows for the
combination of many objects to form an object of higher semantic level where
each part has its own functionality. For example a power distribution network
can be defined from its component parts: poles and towers, transformers,
circuit breakers, switches and power lines.

Figure exemplifying aggregation.
See also:
How
to define a database in SPRING
Conceptual Model
Overview of the Modeling Process
The process of modeling
is an inherent characteristics of the design of image and graphic systems. Four
different universes (abstraction levels) can be distinguished (Gomes and
Velho, 1995):
-
The universe in the real world, which includes the real entities
that will be modeled in the system.
-
The mathematical (conceptual) universe, which includes a
mathematical (formal) definition of the entities that will be included
in the model.
-
The representation universe, where the many formal
entities are mapped into the graphic representations used in the model.
-
The implementation universe, where the data structures and
algorithms for the operations on the geographic data are chosen, based
on considerations like performance, equipment capacity, quantity of data
sets. It is in
this level that coding takes place.
An overview of the modeling process is shown in the figure below.

Phases of the modeling process.
Note that the vision of
modeling just presented is not limited to Geoprocessing Systems. It is
particularly suited to Geoprocessing for it allows the treatment of the
problems of that subject, as can be seen:
-
in the real world universe, we find the data types to be
represented (soil map, urban and rural cadastre, geophysical and
topographic data).
-
in the conceptual universe, we can distinguish among the great
formal geographic data classes (fields and objects), and specialize
these classes in the geographic data types that are commonly used
(thematic and cadastral maps, digital terrain models, satellite
imagery).
-
in the representation universe, the formal entities are
associated to geometric representations that could vary according to
scale, and the chosen cartographic projection. Here we can distinguish
between raster and vector representations, that can be further
specialized.
-
the implementation universe is where the data structures are
chosen and the system is coded.
Based on this vision the
traditional dichotomies of Geoprocessing (object-field and raster-vector) can
be solved, showing that they are found in different levels of abstraction.
This analysis also indicates that the user interface of a GIS should reflect,
as much as possible, the conceptual universe while hiding the details of the
universes of representation and implementation. On the conceptual level the
user deals with concepts that are closer to his reality, and minimizes the
complexity of the different types of graphic representation.
See also:
Relationship between the universes of the
model
Summary
Conceptual
Model
Definition of the Data Model in
SPRING
The geographic database
modeling process in SPRING consists in extending the specialization hierarchy
defined in the model , creating the derived classed of GEOOBJECT, CADASTRAL,
NETWORK, THEMATIC, DIGITAL TERRAIN MODEL and REMOTE SENSOR DATA.
As an example, consider
the following definition of a conceptual scheme for a rural cadastre
geographic database (see the following figure):
-
a class FARM, specialization of GEOOBJECT, that can be further
sub-specialized into LATIFUNDIUM and MINIFUNDIUM.
-
a class SOIL MAP, specialization of THEMATIC, whose instances
store the types of soil of the study areas.
-
a class LANDSAT DATA, specialization of REMOTE SENSOR DATA, whose
instances contain the LANDSAT satellite images over the study region.

Example of Conceptual
Model
The SPRING interface
implements this mechanism of model definition through menus. The procedure to be followed is:
-
define what data types one wants to use in the study (the " categories") and indicate the basic categories of each one;
-
create
a project;
-
create information layers associated to the categories defined in the database, and;
-
edit (digitize or import) the IL's.
See also:
How
to define a Database in SPRING
Notions
on Geoprocessing
About
SPRING
Conceptual
Model
|