| Medial Axis/Ridge Detection [message #30364] |
Thu, 18 April 2002 08:46 |
Ted Cary
Messages: 53
Registered: October 2001
|
Member |
|
|
Hello,
Has anyone written or come across IDL routines that find the medial axis of
a 2D shape? A Google search of the newsgroup for "medial axis" turned up
next to nothing, only one poor post that never became a thread, which is not
promising.
I've started writing a routine myself to do this by finding the "ridges" in
a Euclidean distance map made from a mask of the filled shape. It seemed
like the easiest approach, but I'm having problems optimizing the ridge
detection algorithm which I got from the web. I'm looking for an efficient
way to find "curves" of local maxima in the distance map, hopefully
something clever like what Craig Markwadt and JD came up with last year for
finding local maxima in 1d vectors.
I'm not married to the distance map approach. It was the most intuitive
simply because in distance map pictures the medial axis is so visible to the
human eye, if not to my algorithm (yet). Are there advantages to the
Voronoi cell border approach or to the bi-tangent circle approach that I
should be aware of?
Thanks for any help,
TC
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
