Dear all,
Thanks for your kind replies.
The reason why I can't use the 2D polyfillv is that the points are not
sampled as 2D slices in the rectangular coordinate system. Instead, it
is instead sampled in the spherical coordinate system (every 5 deg
longitudinally) and converted back to rectangular coordinates using
CV_COORD. I am sampling the blobs in the spherical coordinate system
so that I can get the same number of sample points regardless of the
size of the blobs (some are slightly small, some are slightly bigger).
I have checked out IDLanROI. It seems that we can defined a 3D ROI by
loading a (3, n) array. Maybe I can throw in my (3, n) array of sample
points and check whether a point in the 3D space is inside the ROI. In
this way, I might be able to get the points in the volume. But then
again, this method might be quite slow.
Many thanks again for your advices.
Cheers.
On Feb 27, 3:35=A0am, Mike <Michael.Mill...@[EMAIL PROTECTED]
> wrote:
> What if you took your (3,n) values and, for each slice in your data
> volume, find the subset of the n points that fall on (or near enough
> to be considered "on") that slice? =A0Then you could use polyfillv on
> that slice. =A0I haven't done this with quite the data you have, but
> that is the method that I use to create 3D ROIs out of multiple 2D
> ROIs.
>
> Mike
|
