Spatial data structures Flashcards
lecture 7
What are the 2 vector models structures?
- Georelational Model
- Object Based Data Model
What is Georelational Data Model?
Geometry and attributes are stored separately.
Georelational datamodel uses feature identification number to link feature geometry to ite attributes.
The 2 components must share the same ID (syncronised) so that they can ben queried, analysed and displayed in unison.
What is the GDM shapefile?
It is an example if the georelational model.
What are attributes saved as in feature layers?
database .dbf
What is the shapefile?
.shp links all other feature layers to be displayed together.
doesn’t come with topology.
What are the advantages of the GDM shapefile?
- Display faster
- Non-proprietary & interoperable (i.e., can be used across
different software packages)
What are the disadvantages of GDM shapefile?
- Shapes can overlay (not associated with topology)
- Boundaries must be stored twice
- Minimal error checking
- Data quality issues
What is GDM coverage?
Stores vector data - contains both spatial and attribute data (feature classes) for geographic features.
Have topology - determines relationships between features.
Made up of multiple physical files in a folder.
Coverage is proprietary.
What is the Object Orientated Data Model? (OODM)
Stores geometry and attributes in one system.
Geospatial data is treated as objects.
Object = road, symbol, school, co-ordinate system etc.
Property - attributes of features in a layer.
Method - an action that can be performed on a layer
Feature Class - stores spatial features of the same geometry = feature layer.
Feature dataset - stores features that share extent and co-ordinate system - projections are set at this level not at the class level.
How does the OODM differ from the GDM?
- Feature and attributes stored in 1 system.
- Allows spatial features to be associated with a set of properties and methods.
What is a personal geodatabase (OODM)?
- Stores data as tables in a Microsoft access database
with .mdb extension. - Has file limit of 2GB and can only be used in windows
What is a file geodatabase (OODM)\?
- Stores data in many small sized files in a folder with .gdb
extension. - The file geodatabase has no file size limit and can work
across multiple Operating Systems (Microsoft & Linux)
What is topology in the geodatabase?
- Defines topology as relationship rules and lets the user choose the rules,
if any, to be implemented in a vector layer - Topological rules which apply within feature classes is functionally
similar to coverage topology - Rules applied to two or more feature classes are unique to
Geodatabase.
examples:
* Country boundaries must not overlap.
* Country boundaries must not have gaps.
* Contour lines in a vector line layer must not intersect (cross each
other).
What are the advantages of the geodatabase?
❖ Central Storage of files/data
❖ Hierarchical structure is useful for organisation and management
❖When created, new data in a project will be created with already
defined parameters
❖Offers on the fly Topology to features within feature class which
improves data quality
What is the raster model structure?
How raster data is represented.
Refers to the method or format for storing raster data in a computer.
What are the 3 common methods of storing raster data?
- Cell by cell Coding
- Run Length Coding
- Quadtree Coding
What is cell by cell coding?
Presents the simplest raster structure.
stored as a matrix, cell values written into a file by row format.
Functioning at cell level
Ideal for raster surfaces that change continuously and rapidly across space.
e.g. satellite images
What is run length coding?
When cell by cell coding becomes redundant.
Stores values by row and group.
Group = series of adjacent cells with the same value.
for each row, the starting and end cells denote the group’s length.
Run length coding reduces the file size - functioning does not occur at a per-cell level.
Compression and grouping method.
What is quadtree coding?
recursively divides the raster data into quarters.
Once the raster has been divided into 4, each quarter is examined to determine whether all pixels have the same value.
If all pixels do have the same value, then this quarter is not sub-divided
any further.
Where this is not the case, then each of the non homogenous quarters is divided into 4.
Each of these quarters is examined, and those not homogenous are subdivided again.
This recursive subdivision is complete once every ‘quarter’ is homogenous
Raster data compression
- Raster data require considerable memory space
- Data compression refers to the reduction in data volume
- Raster data compression can either be:
1. Lossless
2. Lossy
What is lossless compression?
- Preserve pixel value and allows original raster to be
precisely reconstructed - Desirable for data used for analysis or creation of new
data - Use TIFF or GEOTIFF format NB for MCQ’s
What is lossy data?
- Cannot fully reconstruct original image
- Higher compression ratios
- Desirable for background images, NOT analysis
- Compression can result in loss of geo-referencing info
- Use JPEG format
Rasterisation and integration
- To take advantage of both vector and raster data for a GIS project we
must consider data conversation and integration. - The strength of GIS is it’s ability to do analysis using many different
types an formats of data together. - Satellite Imagery commonly used as a back-drop to vector data –
Google earth - DEM – NB source for deriving contour, slope, aspect data
- Vector data used to verify RS data
- Vector data used as control points for geo-processing
- Vector data used for image stratification – use vector data to divide
image in areas of different characteristics for individual processing. - Big move towards a closer integration of GIS and RS
What are the conversion types?
- Rasterization
- Vectorization
- Hybrid Models (Integration)
Rasterisation
Rasterization converts vector data into raster data
Vectorisation
Converts raster data to vector data
Hybird models
- Use both Vector and Raster models simultaneously
- Most modern GIS are hybrid systems
- ArcGIS (+Spatial Analyst)
- GRASS (limited vector)
- IDRISI (limited vector)
