[Insight-users] one problem of searching stars from sky

[Tao Dai]

Thank you all!

The situation is : some connected pnts are forming one group to be one star;
there're several stars which are not connected with each other.  Any good
suggestions?

Currently I tried kmeans (itkImageKmeansModelEstimator) to cluster them. It
works but one problem is hard to find the right "k" number, since the total
number of stars are unknown.  I definitely will try what Kishore and Lynx
suggested on Monday.


Regards,

Tao

On Sun, Nov 1, 2009 at 1:14 PM, Luis Ibanez <luis.ibanez at kitware.com> wrote:

> Hi Tao,
>
> Your problem looks very interesting.
>
> However,
> It wasn't clear for me, from the description in your email,
> whether the list of bright points are indeed pixels that are
> connected to each other and that form a single star or not.
>
> If they are connected, then, you can certainly use the
> filters that Kishore and Lynx suggested.
>
> However, if you are trying to define a given star by
> aggregating a set of bright pixels that are not
> connected, then you may have to pursue other
> techniques.
>
> Please let us know more details about the problem.
>
> (a screenshot of one of the image will be ideal...)
>
>
>     Thanks
>
>
>             Luis
>
>
> ---------------------------------------------------------------
> On Wed, Oct 28, 2009 at 12:56 PM, Tao Dai <uniquedai at gmail.com> wrote:
> > Dear Luis and ITK list:
> >
> > I am working on one project and trying to do some work like "finding
> > stars":  Assuming there're lots of stars (difference size/brightness) in
> the
> > night sky, I wanna extract those stars and calculate areas and brightness
> > for each.
> >
> > What I am doing first is a binary thresholding
> > (itk:BinaryThresholdImageFilter) and it works well.  However, I don't
> know
> > how to define a cluster of bright points which make a single star. I
> don't
> > know if this make me clear. Let's try another way to clarify: how to know
> > which points are from the same star? any examples available in ITK?
> >
> > Thanks for any input.
> >
> > Tao
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20091101/42eab887/attachment.htm>