Re: Dilation operation on vector

[Terry Duell]

> Hello Petr,

On Sun, 31 Mar 2013 01:43:21 +1100, PetrSt <address@hidden> wrote:

> I'd tried to write something (attached file),

Nothing attached.


> Based on the article I can advice you to check the image processing pkg  
> (not sure by the name), where should be some filters.

I don't think the image processing approach is correct for the problem I  
described.
If I recall correctly, those functions work on a 2D array, and apply the  
filter shape (or motif) in the plane of the image, whereas what I was  
describing is a 1D vector which requires the motif to be applied normal to  
the plane ( if it were an image).

> Also there are some
> functions you could find useful in  30 Geometry
> <http://www.gnu.org/software/octave/doc/interpreter/Geometry.html#Geometry>
> , concerning the convex hull.

As described in the paper I referenced, the convex hull is an approach  
that can be used, but probably not other 3D geometry stuff.

The naive algorithm described in the paper is a simply coded method, and  
appears to work quite well. I have not yet satisfied myself that it  
produces the same result as would be obtained by doing the geometry by  
calculating all the points where the disc is tangent is to the segments of  
the profile, but suspect that any differences will be small.

> I would also advice you to ask about your topic under different keywords
> than dilation, like filtering or morphological filtering, which should  
> attract people more involved

OK, but it really was a particular form of dilation that I wanted to know  
more about, which is, along with erosion, a morphological operation and  
would be in the vocabulary of those in-the-know.
It's not really necessary any more, as the paper I subsequently found and  
referenced provided all the information that I was seeking.

Thanks for your help.

Cheers,
--
Regards,
Terry Duell