Spatial Modeling Continued

1
Spatial Modeling Continued

• Extract and Assign
• Prioritization - Ranking

2
Overview

• Extract and Assign
• Exercise 6 Review
• Prioritization

3
Extract

• extracts summary statistics to either
• a table or
• an attribute values file from an existing image
  file.
• It uses two files,
• an image file and a
• feature definition file.
• The summary statistics in the values file can
  indicate the minimum, maximum, total, average,
  range, or standard deviation of all cells in the
  analyzed image for each identifier in the feature
  definition image.

4
Why Extract

• Data to Information -- We want to add meaning to
  the data set we have through summary statistics
• Remember that we can only have ONE set of data
  per layer!
• Thats the benefit of Extract and Assign

5
Extract Feature Definition

• The feature definition file is an image defining
  distinct features or regions.
• This image must have byte or integer data type
• Nominal, Ordinal, Ratio, or Interval?

6
Extract Image to be processed

• Can be real, byte or integer
• Minimum, maximum, total (sum), average, range,
  population standard deviation, sample standard
  deviation or all listed summary types.

7
Extract Output Table

• Table or Attribute Values File
• Table
• The summary option, all of the above, creates
  tabular output only.
• The table lists the identifier, its legend
  caption (if one exists) and each of the summary
  statistics.

8
Extract Attribute Values File

• The attribute values file output lists the
  identifiers in the feature definition file and
  the summary statistic derived from the image for
  each identifier.
• Now we can use this to do a reclass or ASSIGN!

9
More on Extract - Group

• EXTRACT summarizes all the pixels of the input
  image for each identifier of the feature
  definition image.
• If you have several regions or patches in the
  feature definition image that have the same
  identifier (e.g., separate areas with the same
  land cover class) but you want to generate
  summary statistics for each individual region,
  you must first use GROUP to create a new feature
  definition image in which each contiguous region
  has a unique identifier.

10
Assign

• ASSIGN creates new images or vector files by
  linking the geography of features defined in the
  input file with attributes defined in an
  attribute values file.
• By separating attributes from the geography of
  the features which possess them, this module
  allows one to use the power of spreadsheet and
  database management packages as an integral part
  of the IDRISI system.

11
An Example

• Suppose we have a need to find all soil polygons
  with a average slope of less than, say, 6
• How do we get that?
• Extract and Assign to the rescue!

12
(No Transcript)
13
(No Transcript)
14
Here is Another One

• Suppose we want to find all the parcels which
  have a pixel with a Maximum slope class of less
  than 12 anywhere in the parcel
• Why you would want to do this I dont know but
  just play along!

15
Extract FD Parcels with image-to-be-processed
Slope class with MAX option
16
reclass
Gives
17
Multiply this times Parcels and I have the
parcels that have a max slope class of less than
12
18
Multiply this times Parcels and I have the
parcels that have a max slope class of less than
12
19
So

• Extract to get summary statistics of one image by
  category of another
• or Extract to create a Values Attribute File so
  that we can
• Assign it back (reclass) with the values of
  interest

20
Exercise 6 Overview
21
Basic Steps

• Make the required Boolean images first.
• Note that the parcels image does NOT have unique
  values (IDs) they are owner ids.
• So Group (no diagonals) Parcels first
• I added 1 to the groups so the numbering did not
  start at 0
• Why?

22
Using Extract / Assign Trick
An y parcel with a cell in the buffer will have a
total 1 or more
Road_Buff
GroupedParcel1
Image of area of Parcels which touch the road
buffer
23
Part B

• In part B you are finding building lots of 1 cell
  (165x165) that meet the given criteria.
• You can use Extract / Assign to make map of
  acceptable parcels showing the number of
  buildable lots
• Or use the vector trick to show the lots by
  parcel.

24
Vector Trick???

• Take the Parcel1 Coverage
• Under ReformatgtRaster/Vector select POLYVEC
• Fill in names for both the poly and point files
• A vector file is made from the image

25
Vector layers
26
Prioritization again

• Locating building sites

27
Possible combinations
28
Possible combinations
29
Possible combinations
30
Possible combinations
31
More possible combinations