Comparative Analysis of Cellular Automata and CA Markov in ...

rucksackbulgeAI and Robotics

Dec 1, 2013 (3 years and 4 months ago)

104 views


Presented by:

Paul Lapatka April Vigner Christina Lane Riley Peake


Monday November 24
th

2008

Dr. Suzana Dragićević

Geog 451

Simon Fraser University


Introduction / Objectives


Methods / Procedure


Models


Cellular Automata


CA Markov


Results


References / Acknowledgements


Questions





CA_MARKOV is a combined Cellular Automata , Markov Theory,
Multi
-
Criteria, Multi
-
Objective Land Allocation (MOLA) land cover
prediction procedure that adds an element of spatial contiguity as well
as knowledge of the likely spatial distribution of transitions to Markov
chain analysis.


-
IDRISI Andes



You take two images and input them into the model, it calculates the chances
of one cell changing into another one using several methods and applies for
each iteration. Then a map is produced to suggest the new land
-
use for a
projected time step in the future.



CELLATOM provides for cellular automata in IDRISI. This is typically
used in dynamic modeling where the future state of a pixel depends
upon its current state and that of its neighbors. The rules for changing
states are governed by a filter file and a reclass file. The user specifies
the number of iterations to perform and whether to show the result of
each iteration or just that of the last iteration.



With each iteration, the filter is applied to the image then the resulting
image is reclassified according to the reclass file. This output image is
then used as input for the next iteration.


-
IDRISI Andes



To compare CA Markov to Cellular Automata in the
context of landuse change.



Which modeling approach do you think is better?

1.
Data Preparation


Formatting, resolution,
landuse

classes

2.
Model building

CA: Ca transition rules, reclass file, and filter

Ca_Markov: Editing area file, and probabilities file

3.
Calibration

Changing model parameters to more accurately represent reality

4.
Validation

Using Kappa Statistics, Map Comparison tool kit


*Referencing Academic articles used during all processes

3 Cell

Neighbourhood

Actual 2001 Landuse

Cellular Automata Projected


2001 Landuse

3 Cell Neighbourhood

4 Cell Neighbourhood

Actual Landuse 2001

Cellular Automata Projected
2001 Landuse

-

Static

-
“Black box”

-

Every land use changes
(including water)


CA Markov Projected

2001 Land
-
use

Actual 2001 Land
-
use

* Results not accurate Negative growth suggested by CA Markov *

Cellular Automata

CA Markov

Our results show that CA outputs more closely resemble actual data than
outputs from CA Markov.


Actual Change from 1996
-
2001

4 Cell Neighbourhood CA Change

from 1996
-

2001

3 Cell Neighbourhood CA Change

from 1996
-
2001

CA Markov growth from 1996
-
2001


IDRISI Andes Help Files


Chris Bone


Suzana Dragićević


Statistics Canada
. Date of Posting: February 1
st

2007 01:35:01 PM. Statistics
Canada. Last accessed on November 22, 2008
[http://www12.statcan.ca/english/profil01/CP01/Details/Page.cfm?Lang=E&Geo1=CMA&Code1=933__
&Geo2=PR&Code2=59&Data=Count&SearchText=Vancouver&SearchType=Begins&SearchPR=01&B1=
All&GeoLevel=&GeoCode=933
]



End