I forgot: If there is someone who can help me adjust this analysis to analyze imagery data, please contact me. I know almost nothing on analysis of such data but it seems it could be useful for it.

Jarosław Jasiewicz pisze:
Hi list!

I added new module r.fuzzy.system to svn repository:

this is standalone classification module, to work require only sets and rules definition (see module description and small tutorial in description.html)

what module do (internally):

- parse sets and rules text files before calculation
- creates fuzzy sets based on user's definition
- implicate fuzzy rules form antecedents to consequent based on user rules
- aggregate consequents
- defuzzify final set

additional features:
- output fuzzy maps for particular rules
- output consequents and aggregated fuzzy set for particular cells (without map calculation)

In spite that module has the same name as that created by Francois Declaux, it is completely new tools without any references to that except it was created for that same reasons. Module r.fuzzy system do not require r.fuzzy and r.fuzzy.logic to work, but these modules can be used to trace intermediate results (they use the same fuzzy algorithms)

It fact module do almost all what I assumed, but there are probably lot of bugs and errors. Because it was rather difficult coding (especially text file and expression parser) any help with testing will be appreciated. I will also appreciate if someone with better English than mine can help me write tutorial on fuzzy logic in GIS

Known issues:
If there are cells for which membership in all sets is 0 module returns null. It is probably difficult to solve without extensive graphical UI with additional correction and check-out tools so probably will remain unsolved.

regards
Jarek Jasiewicz
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user



_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user

Reply via email to