Best fit routine

I have a 30 sided polygon that I want to best fit inside another 30 sided
polygon that is similiar to the first but not exact. I would like for each
of the vertices to be as close to each other as possible, other words, the
minumum average coordinate error. With polygons of this size this may take
some type of perhaps least squares routine. I need something that will be
a mathmatical solution.

Any suggestions or thoughts appreciated.

 

Are these regular polygons (all sides and angles equal) or what?
___

 

No, they are irregular in shape. They look similiar, but the lengths of
sides and interior angles are subtly different.

 

If you convert your polyline to a region you can extract the geometric centre (centroid) of the shape as a co-ordinate. Run _massprop on the region and you'll find this info in the text window. You can then match the centroid of each shape, giving a best fit.

 

Wellll... yes, but, that may not actually give the best fit. the polygon
that is being translated may also need rotated. matching centroid is a
start, but again that may not really give the best fit in as much as the
polygon may need to shift centroid away from centroid, or so it seems. I am
not a math wizard, made c's and it's been awhile. Matching centroids is one
thing but making the vertices match with least amount of coordinate
difference may be another.

Your thoughts?

centre (centroid) of the shape as a co-ordinate. Run _massprop on the region
and you'll find this info in the text window. You can then match the
centroid of each shape, giving a best fit.

 

lsp was removed by OE..... who is the OE

 

OE - outlook express.....

Matt

 

oh,... duh

Found the Helmert .vlx routine, very cool, may be what I am looking for