GIS: Datum and projections / GIS TRAINING

All the coordinates of objects in the same database must be expressed in the same system in order to be comparable

YOU MIGHT ALSO LIKE: General Principles Of Gis/Gis: Definitions And Main Functions

---

Projection: represent an ellipsoid on a plane

• A cartographic projection is a mathematical operation allowing to represent a portion of the ellipsoid on a plane, by estimating the deformations induced by this operation on curvilinear distances, angles, directions, curvilinear surfaces...
  

---

Databases

From the object to the collection of objects: databases

• The purpose of the schematization of reality here is to describe not a single object, but a set of objects. It is therefore even more reductive since the context is that of a collection. Attributes must, therefore, be common to all objects in the collection.

• A database is an association between a schematization of reality and the objects describing reality according to this schema.

• The need for computer management is obvious, to manage all objects in relation to descriptors (attributes), to manage the links between objects and to connect objects to each other. This is managed by a database management system (DBMS).

Database management systems: objectives

• Physical and logical independence between data and application programs

• Object persistence

• Centralized data administration

• Optimal management of computer memory and efficient data access

• Sharing data between users and managing competing accesses

• Reliability, integrity, and consistency of data

• Data Security

• Interactive queries, declarative data consultation, access to non-computer specialists

---

The relational model

  -A simple description model

• Objects are described only by simple type attributes (for example, no attribute in R2 or R3, no recursive definition, no methods)

• The set of objects described by the same attributes is called a relationship. Objects are called tuples

• All the objects in a relationship can be represented by a table (line=tuple, column = attribute)

The most common DBMS are relational: ACCESS, DBASE, MySQL, ORACLE, etc.

-The tuples are manipulated using relational algebra operators, a formalism that allows the content of the database to be queried:

• union
• Cartesian product
• projection
• selection
• joint

Relational algebra allows queries to be expressed by a sequence of operators. Expressed in a high-level language, the query ensures the objective of physical independence between data and application program (SQL languages).

---

The extension of the relational model

• Simple and powerful, the relational model only handles data related to the natural order (in the selection and join criteria). It does not allow to properly process data of dimension 2 or more, such as location.

• To treat the location of geographical objects with a relational system, it is, therefore, necessary to extend the relational model and algebra for R2 or R3 data. The operations related to this type of data are based on the distance between objects, on ensemblist notions (union, intersection, belonging), on topological notions (adjacency, connectedness), and no longer on a simple order relationship.

Implementation in GIS software

  -Management based on descriptive joins

• Geometry and topology are stored in separate files

• Descriptive information is managed by a traditional DBMS

• A unique identifier per object makes the link between geometry and description

• Examples:

            -Arc view
            -ArcInfo

  -Management based on an extension of the relationship to localization

• Geometry and description are managed together

• Two-dimensional indexing is possible

• A two-dimensional BD motor (SDE) is used

• Examples:

           -Sav GIS
           -Geodatabase (ArcGIS)
           -Spatialware (MapInfo)

YOU MIGHT ALSO LIKE: What is ArcGIS?: ArcView | ArcEditor | ArcInfo

---

Limitations of the extended model

• Hierarchy, belonging, classification by neighborhood... need to introduce methods and types of complex objects (zones, lines, points), hence the de facto extension of the relational model towards object-oriented concepts

• Query or consultation operations take the form of real methods and call into question the principle of DBMS: GIS becomes a real application program, and cannot be limited to data management.

A look back at the spatial integrity constraints

• Some geometric or topological constraints will always have to be checked (for example, an area must always be closed), as they depend on the logic model or internal description model

• Others depend on the semantic definition of the collection (a road network must always be connected, but a telephone network may not be)
  

  -Geometric constraints on arcs

• Simplicity (overlapping of an arc on itself)
• Extra-simplicity (intersection or duplication of arcs)           
• Inclusion
• Closing up
• Connectivity

  -Topological constraints of type (area, line, point)

• Closing of areas
• Belonging of the centroid to its zone
• Connection of zones or networks

  -Relationship constraints

• Unique key constraint
• The constraint of belonging to a domain
• Neighborhood constraint
• Metric constraints

  -Joint constraints

• Geometric joining constraints: geometric belonging, inclusion (junction of boundaries and hierarchy of relationships), sharing (sharing of arcs between collections),

• Descriptive join constraints