geographic information / GIS TRAINING

Knowledge: 

• How to understand and represent reality to deal with a computer?
• universal vision or contextual vision?
• How to define criteria description of reality without a specific problem, to begin with?
• Precision, scale, and description, modeling reality: the approach of the geographer.

• The real world (no abstraction)

• The conceptual model (conceptual modeling of reality)

• The logic model (organization model of computer-related)

• The physical model (an internal organization to the application)

The geographic data

• Recording measurements were taken at a certain place at a certain time in the real world

• Combines location, time, and descriptive attributes

• Difficult to handle in conventional systems data management, which are not equipped to higher dimensional data 1

The spatial object

• An object in information theory is an encapsulated set of attributes and methods for describing knowledge and behavior to the contextual view of reality.

• Spatial Object has three major components: location, description, behavior
  

The different types of attributes of a geographical object:

   Information descriptive :

Limitations of the mapping model

   The limits of geometry and cartographic model:

• It is assumed that classical geometry can be used to describe the location of geographical objects. We, therefore, introduce discontinuities into reality by using zone, line, point schematization to define geographical objects. Accuracy or uncertainty are not addressed by these description models. Space is not treated continuously, the geographical definition of objects is discontinuous and greatly simplifies reality.
• Is the geometric description in zone, line, point sufficient to describe geographical objects in a satisfactory way?

Cartographic model and computing power

   The contributions of information technology:

• From geographical description to computer description: should computer science take over the geometry of objects or question the overly simplistic schematization of the cartographic model? Will it make it possible to improve space modeling, whereas for the moment it only uses existing schemas (the cartographic model)?
• Aren't the treatments applied to geographical objects in GIS too sophisticated in relation to the validity of the schematization of reality?

From geometry to computer science: a logic model

• The localization of objects is totally simplified (all objects have the same size and shape). Only one object geometry (the mesh) is defined to which all descriptive attributes are attached.
  
• raster data fills the entire space 
• The accuracy is fixed once and for all, and generally degrades the accuracy of the cartographic model from which the information is derived 
• The implementation of computer algorithms for analysis operations is easy, but the difficulty of spatial analysis remains complete 
• A distinction must be made between "raster representation" and "raster-type" information, such as satellite images or scanned photographs: the accuracy depends on the sensor and does not degrade the original data.

Vector representation

• Zones and contours, networks and lines, points. The geometric definition of the objects in the cartographic model is retained, but a mathematical description (in R2 or R3) is changed to a simple computer description in a discrete set (with a finite number of parameters):
• Representation of an arc by a finite set of points. Representation of an area by a set of arcs. Graphs and networks.
  
• The objects of the conceptual model are not modified, the geometric accuracy is maintained, the graphical-descriptive relationship is not disrupted
• Storage space is low
• The structure allows two-dimensional indexing
  

     Pixel representation

• Satellite images or scanned aerial photographs are different from a raster representation: the grid location comes directly from the sensor. Information is the value of a cell, called a pixel. It is not a logic model, but a conceptual model of reality description (model defined by the sensor manufacturer, who chooses the resolution, wavelengths).
  
• The main purpose of remote sensing is to move from pixels (and the descriptive values it contains radiometry or grayscale) to the location of objects defined by their descriptive content (land use, vegetation type, etc.) by grouping pixels.
• Another approach is to treat the pixel as an area object. Everything depends on the size of the pixel in relation to the definition of the geographical object being studied (can the pixel be assimilated to a point in the mathematical space or should it be considered as an area?