The Raster GIS basics

1
The Raster GIS basics

• From Data to Information

2
Today --

• Layers how GIS data is organized
• The Raster some details
• Digitizing
• Coding rules
• A Demo and the manuscript data

3
GIS data

• Geographic data where is it
• Projection
• Datum
• Attribute data what is it
• Metadata where did it come from

4
GIS LAYERS
Geographic location (X,Y)
Streams
Power lines
Landuse
Roads
5
A little more on Raster layers

• In IDRISI layers are Maps -- but sometimes I
  will call them images
• A better word is theme since each map can only
  represent one theme or part of a theme. BUT
  IDRISI DOES NOT USE THAT TERM
• In 356/556 map layers can be photographic or
  satellite images or raster data images (maps).
• Although in general images mean photographic or
  satellite data.

These can be called dumb rasters since color is
the only info in them
6
The RASTER of raster systems

• A raster is a
• table like array of values
• a grid of cells or pixels

7
The RASTER of raster systems

• A raster is a
• table like array of values
• a grid of cells or pixels
• This is a 6x5 (C,R) ARRAY
• Cells have attributes

Symbolization of the attribute values
8
The Raster

• The Geographic position within the raster is
  determined by the cell coordinates in the raster
• The Geographic position in the real world is
  determined by the coordinates of the entire
  raster in real world coordinates
• The attribute data is the single numeric value
  stored in the cell
• The metadata is in the document file (.rst)
  associated with the image data (image, raster,
  array all the same thing).

9
Coordinates

• The Geographic position within the raster is
  determined by the cell coordinates in the raster

• IDRISI uses the mathematician's array numbering
  it starts with 0
• This can be a booby trap since the last cell in
  the second row has coordinates (r,c) 1,5
• Not 2,6!

10
Coordinates

• The Geographic position within the raster is
  determined by the cell coordinates in the raster

Max x, Max y
Min x, Min y
11
Cell Resolution

• a.k.a. spatial resolution or RESEL
• The smallest object/feature that is mapped
• To be very accurate your pixel size should be
  smaller than your smallest feature.
• But you run into the tradeoff between small size
  and
• a) storage requirements and
• b) processing time
• NIFKIN-- size of each pixel or cell is 165 ft x
  165 ft

12
Attribute Data

• What is it? - Descriptive Data
• In Vector GIS a huge amount of attribute data can
  be attached to a map feature
• In Raster GIS attribute data is contained within
  the cell or pixel one value per cell
• Numeric attributes!

And Only Numeric!
13
Attribute Data Categories

• Data types
• Nominal
• Ordinal
• Interval
• Ratio
• When we do analysis -- Boolean (1,0) later....

14
Nominal data

• Numbers assigned have no meaning whatever
• They are just nominal (named) assignments
• Like Arnot Soil is a 6, Lordstown is a 9
• You cant do any mathematics with nominal data
• What is Lordstown Arnot soil?????
• 15???? Does not mean anything!
• http//www.cmh.edu/stats/definitions/nominal.htm

15
Nominal Data Examples
Forest Type
16
Ordinal data

• Ordinal attribute data does have some meaning but
  intervals between numbers have no meaning
• Road types
• 1 Interstates
• 2 State Highways
• 3 County roads
• The above is Class data
• Other data types can be changed into
  ClassesElevation can be classed into ranges of
  elevation
• http//www.cmh.edu/stats/definitions/ordinal.htm

17
Ordinal Data
Soils
Very Good
Good
moderate
Poor
18
Interval data

• Like Ordinal but now the interval has meaning but
  there is NO natural zero it is continuous data.
• For example Temperature
• The intervals have the same meaning no matter
  where you are on the scale
• There is, however, for F or C, no natural zero
• Absolute temperature (Kelvin) DOES have a natural
  zero (where all atomic motion stops)
• Cant use for ratios. Is 80F twice as warm as
  40F?
• http//www.cmh.edu/stats/definitions/interval.htm

19
Interval Data
pH
4.5 5.4
5.5 6.4
6.5 7.4
7.5 -8.4
20
Ratio data

• Like Interval except now there is a natural zero
  point.
• Temperature in degrees Kelvin
• Rainfall
• Weight (80lbs IS twice 40lbs!)
• Elevation above mean sea level
• www.cmh.edu/stats/definitions/ratio.htm

21
Ratio Data
Precipitation, in
0.0 10.0
10.1 20.0
20.1 30.1
40.1 50.0
22
Note

• You really cant tell what kind of data it is
  just by looking at the layer
• You have to know which kinds of data can be
  mathematically processed in a GIS

23
Digitizing!
24
Digitizing - Creating Digital Data

• One of the skills youll take home from this
  course
• It is important for understanding data creation
• And it improves your understanding of rasters
• That is why you are doing it by hand in Exercise
  2!
• The basic operation here is to put numeric data
  into cells in the raster

25
Digitizing - First Step

• Plan!
• You have to decide how you are going to enter
  your data.
• In the future, you will do more planning to solve
  problems
• Logbooks should have a proposed spatial model!

26
What should you think about?

• Data Type
• Nominal, ordinal, ratio, interval
• Classification system -- Legend
• What do you want to do with the data?
• Data Entry -- Coding Rules

27
Coding rules
Or how you decide what value to put into any
given cell

• Point What value under the point
• Majority Most common value, gt50
• Average Average
• Prominence Most important, presence or absence
  of something
• Connective Need to connect area of same value

28
Exercise 2 and 3

• In Exercise 2 you will create paper versions of 3
  IDRISI maps (images)
• In Exercise 3 (next week) you will use Excel to
  put the data into digital form.
• You will e-mail your elevation map layer
• I will then compute the class average and
  standard deviation of the data so you can see
  where you had differences of opinion.

29
Elevation on paper
30
Soils in IDRISI
31
Soils on paper
If Wp very important is this one Wp or Vp?
32
Summary

• Raster data is like a spread sheet cells
• Raster data can only have 1 attribute
• Although I have never seen students try to do
  math with nominal or interval data it is good to
  understand the differences
• Go forth and conquer

33

?
34
Data Storage

• You cannot save data on the cluster machines!
• The Nifkin database is 129K
• So you will need a flash memory stick of at least
  512K