[net.physics] 2-d spline

peshkin@fas.ri.cmu.edu (Michael Peshkin) (12/06/85)

---
I need a spline subroutine (any language) for smoothing a function of
two variables f(x,y).  I have f(x,y) for discrete data points (x,y); a
spline allows interpolation between the discrete points.  A
non-interpolating, or "smoothing" spline does not insist that the values it
provides agree exactly with the data at the discrete data points.  It is
content to miss the points (a little) so that the curve it produces isn't 
too ragged.

If you have, or know of, such a beast please let me know.  

Michael Peshkin       (phone (412) 578-3730)
Robotics Institute  Carnegie-Mellon Univ.  Pittsburgh, PA 15213
---------

gwyn@brl-tgr.ARPA (Doug Gwyn <gwyn>) (12/06/85)

> I need a spline subroutine (any language) for smoothing a function of
> two variables f(x,y).

DeBoor's "A Practical Guide to Splines" has such an algorithm,
which IMSL lifted and called something like ISSCU.  This code
does not work well unless the data are properly scaled.  I'll
mail the requestor a copy of my C code that implements this,
including scaling, assuming I can figure out his mail address.