[SciPy-user] neighbourhood of randomly scattered points [Was: choose at random all elements in a array...]

fred fredmfp@gmail....
Tue Aug 28 02:25:30 CDT 2007


fred a écrit :
> Hi all,
>
> I would like to show how I use this,
> with a result, which seems to be strange to me.
>
> I have a float array, with dimensions of 501x501.
> The values in this array are the scalar values at the point (x=i*dx, 
> y=j*dy,)
> with i=0..500 & j=0..500.
>
> I flatten this array and get 15000 random (with random.permuation 
> method) values from this array
> and thus get 15000 points.
> Then I want to find out the number of neighboors from each of theses 
> points,
> in a neighboorhood of (2*20+1)x(2*20+1).
>
> Thus, I get an array with coords and the number of neighboors.
> x0 y0 nb0
> x1 y1 nb1
> ...
>
> I display the result with mayavi2, and get the following snapshot:
> http://fredantispam.free.fr/snapshot.png
>
> Structures clearly appear.
> (I don't consider side effects, of course).
>
> What could be the meaning of this ?
> If it was "totally random", should I not get no structure at all, 
> something like a white noise ?
>
> What am I doing wrong ?
>
> How can I get a totally random scatter points ?
No one can help me on this issue ?

TIA.

Cheers,

-- 
http://scipy.org/FredericPetit



More information about the SciPy-user mailing list