Spatial Data Formats
Post on 22-Feb-2016
29 Views
Preview:
DESCRIPTION
Transcript
CS 128/ES 228 - Lecture 4b 1
Spatial Data Formats
CS 128/ES 228 - Lecture 4b 2
Need data for a GIS?
• Just photograph a topographic map
• Better yet, download one from the internet
• But are the roads, buildings, and other “objects” on this photo GIS layers?
CS 128/ES 228 - Lecture 4b 3
Stages of development:1. Conceptual model: select the features of
reality to be modeled and decide what entities will represent them
2. Spatial data model: select a format that will represent the model entities
3. Spatial data structure: decide how to code the entities in the model’s data files
CS 128/ES 228 - Lecture 4b 4
2. Spatial data models
1. Raster
2. Vector
3. Object-oriented
and… attribute data
Spatial data formats:
Fig. 3.1 in 3rd ed.
CS 128/ES 228 - Lecture 4b 5
Raster format Features
represented by cell contents
Spatial precision limited by cell size
Surfaces modeled as continuous values (almost)
Fig. 3.9 in 3rd ed.
CS 128/ES 228 - Lecture 4b 6
Vector format Discrete features
explicitly represented
Spatial precision limited by number format
Surfaces shown by contours rather than continuous values
Fig. 3.9 in 3rd ed.
CS 128/ES 228 - Lecture 4b 7
Object-oriented formats
Leave details for CS majors
Fig. 4.17 in 3rd ed.
CS 128/ES 228 - Lecture 4b 8
Thematic data (a.k.a. “attribute data”)
Quantitative or descriptive
May represent 1 or many themes
Tied to a spatial reference
Represented differently in raster vs. vector formats
CS 128/ES 228 - Lecture 4b 9
Scales of measurement
Data Unit ScaleResort name text Nominal
Resort ranking value Ordinal
Winter temp. oC Interval
Size of ski area m2 Ratio
Heywood et. al. 2006 – Table 2.1
CS 128/ES 228 - Lecture 4b 10
Spatial modeling in raster format Basic entity is the
cell
Region represented by a tiling of cells
Cell size = resolution
Attribute data linked to individual cells
CS 128/ES 228 - Lecture 4b 11
Attribute data in raster formatAttribute data are used to create symbology for
each cell
CS 128/ES 228 - Lecture 4b 12
Additional attribute data Some GISs provide a VAT linked to individual
cells (e.g. ArcInfo GRID)
VAT data then accessible to database management system
Unlimited additional fields
CS 128/ES 228 - Lecture 4b 13
Attribute data for a vector layer Each entity is linked to a row in an attribute
table
Themes not (usually) displayed but available via Identify tool
CS 128/ES 228 - Lecture 4a 14
Vectors are good at modeling …
roadways or wiring consist of discrete components; types and order of the connections are key
Spaces between the network components generally not of interest Bottom : http://www.dunereview.com/electricalupgrade-1.htm
… networks
CS 128/ES 228 - Lecture 4a 15
Rasters are good at modeling …
they model a continuous feature as a 2- or 3-D layer
every location has a value, even if only interpolated from discrete samples
Both: http://snobear.colorado.edu/Markw/Research/ESRI/ESRI.html
… surfaces
CS 128/ES 228 - Lecture 4a 16
Topographic maps use contours…
…but the elevation between contour lines is undefined
CS 128/ES 228 - Lecture 4a 17
Digital terrain models
Every cell has an elevation value
Fig. 3.32 in 3rd ed.
CS 128/ES 228 - Lecture 4a 18
Precision agriculture
Aerial photograph Soil pH Crop yield
CS 128/ES 228 - Lecture 4a 19
Oceanography
Estimate of phytoplankton distribution in the surface ocean: global composite image of surface chlorophyll a concentration (mg m-3) estimated from SeaWiFS data (Source: NASA Goddard Space Flight
Center, Maryland, USA and ORBIMAGE, Virginia, USA).
CS 128/ES 228 - Lecture 4b 20
Rasters are a type of Tesselation A closed shape or polygon that repeats on all
sides without any gaps or overlaps
Three regular polygons tesselate the plane:
Square Equilateral triangle Hexagon
CS 128/ES 228 - Lecture 4b 21
TilingsIn 1922 Escher visited the Alhambra palace and saw the wall tilings of the Moors. He was excited to find other artists who had been captivated by tilings, but also made this revealing comment: "What a pity their religion forbade them to make graven images."
CS 128/ES 228 - Lecture 4b 22
Escher’s “tesselations”
CS 128/ES 228 - Lecture 4b 23
Quilters also tesselate
Designing Tesselations by Jinny Beyer
CS 128/ES 228 - Lecture 4b 24
Effects of resolution – rasterLarger cells: less precise
spatial fix
line + boundary thickening
features too close overlap - less detail possible
Fig. 3.10 in 3rd ed.
CS 128/ES 228 - Lecture 4b 25
Advantages of raster format many data sets
available
easy to overlay multiple themes
able to represent multiple continuous surfaces
different file formats readily inter-converted
fast computer lookup and display
CS 128/ES 228 - Lecture 4b 26
Limitations of raster format poor representation of
discrete objects
exact boundary location difficult
constant resolution throughout the region modeled
generates very large data sets
difficult to change projection or coordinate system
CS 128/ES 228 - Lecture 4b 27
Raster layers don’t share well
CS 128/ES 228 - Lecture 4b 28
Raster layers are normally projected•Note the datum and projection/ coordinate system
• Special software needed to re-project
CS 128/ES 228 - Lecture 4b 29
Summary: Raster format Huge amounts of
spatial data are available in raster format
Rasters are the format of choice for continuous features
Rasters do a poor job of representing discrete features
top related