Xref: utzoo sci.math:17845 comp.graphics:18392
Path: utzoo!utgpu!cunews!cognos!geovision!mark
From: mark@geovision.gvc.com (Mark Baker)
Newsgroups: sci.math,comp.graphics
Subject: Best fit parabolas
Message-ID: <1605@geovision.gvc.com>
Date: 3 Jun 91 19:02:50 GMT
Reply-To: mbaker@geovision.gvc.com (Mark Baker)
Organization: GeoVision Corp, Ottawa, Canada
Lines: 30


As part of a project I'm currently working on, we are building
a series of utility routines that will give a 'best fit' conic
given a set of input points.  Currently we have algorithms that
produce good quality cirles, cirular arcs, ellipses and elliptical
arcs.  However, we're having a terrible time producing good quality
'best fit' parabolas.

For all of these conics we're using a least squares approach that
is able to give us nice 'average' curves given a set of input points.

Does anyone know of any non-proprietary algorithms that would be
able to produce the desired result?

Please respond by e-mail as I don't scan these groups on a regular
basis.  If we achieve a solution with acceptable results, I'll post it
to these groups.

Thanks,
Mark.

------------------------------------------------------------------------------
You think.                                          \  Mark E. B. Baker
You wink.                                            \  GeoVision Corp
You do a double blink.                                \  Ottawa, ON, Canada
You close your eyes and jump. - Bert the chimney sweep \  (613) 722-9518 x205
------------------------------------------------------------------------------
Disclaimer:  If I want you to have an opinion,         / mbaker@gvc.com
             I'll tell you what it is. - J.G. Cramer  / uunet!geovision!mbaker
------------------------------------------------------------------------------