[GRASS-user] errors in r.le.patch calculations

[Hamish]

vaniscotte wrote:
> When I am runing "r.le.patch" I meet systematically some erroneous 
> calculations for certain columns of pixels.
> This happens while I am using the "moving wondow" calculation for some
> attributes  (class of metrics: "a"). Whatever the metric calculated
> some vertical bands of the output image are always and abnormaly set
> to 0. The problem still arises when I try several moving window
> parameters  (radius or type) and several version of grass (grass60 or
> grass61)

no solution, but I recommend you try r.le.patch from GRASS 6.2.1, as
there have been some recent fixes there which may solve your problem.
(or not)

> Has someone ever met this problem? Is it a simple display problem or a
> bug in the imagery calculation?

d.what.rast or d.rast.num might help answer that.

Also note that older versions of GRASS treated NULL raster values as
"0", and unknown color rules will default to white in the display.


Hamish