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:

  1. define what data types one wants to use in the study (the " categories") and indicate the basic categories of each one;
  2. create a project;
  3. create information layers associated to the categories defined in the database, and;
  4. edit (digitize or import) the IL's.


See also:
How to define a Database in SPRING
Notions on Geoprocessing
About SPRING

Conceptual Model