Re: Need help in curve fitting
- From: "John D'Errico" <woodchips@xxxxxxxxxxxxxxxx>
- Date: Thu, 29 Jun 2006 16:16:23 -0400
Sumit Tandon wrote:
Thanks for the help Duane and John,
I am trying it out. I just have to keep an eye on the error.
I had one more query...what happens if the points dont lie on the
periphery of an ellipse/circle....then is there any way of getting
an
equation which for the closed loop on which they lie ? ...fitting
an
ellipse in that case may result in a large error...
i have heard of NURBS....but i dont have any experience in
that...does it give an equation as an output, to represent the
collection of points....if so, is there any MATLAB code available
online that i can use for simulation...
The problem is, if its not an ellipse,
then what is it? There are an infinite
number of model forms that you might
pose. Finding an "equation" that
represents the curve is not always
a simple task.
If you can find the "center" of the
curve, then perhaps you might convert
to polar coordinates. Form a model for
r(theta), perhaps using a least squares
spline. Given that you use a spline,
the actual origin for this coordinate
system need not be perfectly chosen.
John
.
- References:
- Need help in curve fitting
- From: Sumit Tandon
- Re: Need help in curve fitting
- From: John D'Errico
- Re: Need help in curve fitting
- From: Sumit Tandon
- Need help in curve fitting
- Prev by Date: Simulink Model Parameter Query Tool ?
- Next by Date: Substitution in matrix
- Previous by thread: Re: Need help in curve fitting
- Next by thread: CompIMAGE2006 - Remainder: SUBMISSION PERIOD ENDS TOMORROW
- Index(es):
Relevant Pages
|