Dear Tim

Thanks for your reply. I'm one of the coaches of this thesis.

> Right - one could probably apply blur effects and similar to lines to
> represent their fuzzyness - Qt4 already has some infrastructure for this
> kind of thing e.g.
> http://doc.qt.nokia.com/4.6/qgraphicsdropshadoweffect.html#details

=> Is this also accessible through Python (plugin)?

> I'm not sure about using two separate layers as you describe above - it
> might be nicer to make fuzziness simply an attribute of the feature -
> though I guess it doesnt cope well for the scenario where individual
> segments of a feature have different fuzziness levels...

In fact, the fuzziness attribute has to be somehow 'attached' to segments.
In the use cases I know often there are some segments along the
boundary which are not fuzzy but crisp on purpose.
An example might be where a street touches the fuzzy feature.
So since segments of a polygon can't be accessed from 'outside' one
solution might be to have a "shadow" layer which consists of
polylines.

=> Do you have another idea on how to implement such "line segment
atttributes" in QGIS?

Yours, S.


2010/9/27 Tim Sutton <[email protected]>:
> Hi
>
> On Wed, 2010-09-22 at 17:47 +0200, Annina Hirschi Wyss wrote:
>> Hello,
>>
>> I'm preparing my master's thesis on the subject of handling fuzzy objects 
>> using the fuzzy logic, with the focus on geometries (not on the thematic 
>> level).
>> I wanted to know, if there has already been a discussion about implementing 
>> this in QGIS or in a plugin.
>>
>
> As far as I know, no one is working on this or has discussed it. But its
> an interesting topic and it would be nice to see some real innocation
> happening in QGIS.
>
>> Showing the uncertainty of boundaries is needed in some cases of spatial 
>> planning, when the "crisp" geometries are not yet defined.
>>
>> A possible use case could also be to draw a polygon with certain attributes, 
>> then the border of the polygon "copied" to one or more oriented polyline 
>> layers with special attributes (e. g. the distance of the uncertainty, to 
>> which the fuzzy-function is applied). These polylines are linked by a 
>> foreign key to the polygon-feature. This would certainly require the use of 
>> a database, namely Postgre/PostGIS.
>> A polyline layer (linked to the polygon) can also be "crisp" e.g. if the 
>> border is a street or a river. So it should be possible to combine crisp and 
>> fuzzy segments for the same polygon.
>> Visualising this could result in a polygon with a sort of "thick pen-trace"  
>> (the thickness corresponding to the distance-attribute of the polyline) 
>> boundary, showing the uncertain band. The thickness corresponds to the 
>> distance-attribute of the polyline.
>>
>> Another possible visualisation (sort of fuzzy buffers) is presented in the 
>> following paper: http://www.geo3.nl/12%20fuzzy%20set%20theory.pdf (fig. 1b).
>>
>
> Right - one could probably apply blur effects and similar to lines to
> represent their fuzzyness - Qt4 already has some infrastructure for this
> kind of thing e.g.
> http://doc.qt.nokia.com/4.6/qgraphicsdropshadoweffect.html#details
>
> I'm not sure about using two separate layers as you describe above - it
> might be nicer to make fuzziness simply an attribute of the feature -
> though I guess it doesnt cope well for the scenario where individual
> segments of a feature have different fuzziness levels...
>
>
>> If there are developers interested in this, I would be glad to know, as my 
>> own programming skills are rather basic.
>
>
> Look forward to seeing your proof of concepts!
>
> Regards
>
> Tim
>
>>
>> Thanks in advance
>> Annina_______________________________________________
>> Qgis-developer mailing list
>> [email protected]
>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
>
> _______________________________________________
> Qgis-developer mailing list
> [email protected]
> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
>
_______________________________________________
Qgis-developer mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/qgis-developer

Reply via email to