[GRASS-user] Re: count pixels by attribute in sampling units

[Patrick Giraudoux]

Sorry to reply to myself... I just though this could help other listers...

It seems that r.le.patch provides a way to get a count of cells by 
attribute in each sampling unit with the option att=a5. One must 
previously define each attribute value as a group (e.g. with r.le.setup) 
and things work fine at least with rectangles. I will have some control 
trials with circles later on...

Patrick


Patrick Giraudoux wrote:
> Dear listers,
>
> I am exploring r.le and r.li, but I am afraid those two sets of 
> programmes are providing what I need only partly. Before using them 
> for shape and diversity analysis, I just would like to get the number 
> of pixels (or the area) by attributes types (eg the attribute is a 
> class of CORINE landcover or any classified image) in each sampling 
> unit into a text file. Sampling units would be circles of a given 
> radius centered over sites (overlayed on the raste to analyse). I can 
> provide either a vector file with the center points or imported 
> polygon shapefiles. Surprisingly it seems that this simple statistics 
> is not provided in r.le (maybe wrong ?). I suppose this may be due to 
> the fact there is a much more simple way to do it in GRASS, but cannot 
> find it in Neteler and Mitasova book.
>
> Any hint welcome,
>
> Patrick
>
>
>