[comp.graphics] Intermediate Control Points for Bezier Curve.

as@progress.COM (Amitabh Shukla) (10/23/90)

Given a free form curve ( an array of points ), how do we calculate the two 
intermediate control points? I was going through the bible (Foley Van Dam)
and realized that it tells you how to draw curve when you have the 4 
control points but it is not easy to understand how to go the otherway.

I would appreciate if someone can guide me towards a solution. 

					Amitabh Shukla

email  address:  			as@pscdev.progress.COM

joerg@cat.uucp (Markus Schichtel) (10/27/90)

Amitabh  Shukla (as@progress.COM) schrieb am 22.10.:
> Given a free form curve ( an array of points ), how do we calculate the two
> intermediate control points? I was going through the bible (Foley Van Dam)
> and realized that it tells you how to draw curve when you have the 4
> control points but it is not easy to understand how to go the otherway.
>
> I would appreciate if someone can guide me towards a solution.
>
>                                       Amitabh Shukla
>
> email  address:                       as@pscdev.progress.COM

Frankly your question is not that precise,But since you are talking of four
control points I assume you use a cubic Bezier curve,in which case the theory
tells you that the first two resp. last two controls determine the tangent to
the the curve at the endpoints.In other words if you prescribe the tangents at
the endpoints then you have also the two intermediate control points.This
property follows from the Casteljau Algorithm which is commonly used to
calculate points on the curve.Presumably you want to approximate an given
array of points.In that case it is not at all necessary to compute the missing
controls.You just make them the parameters in your approximation process.
Standard method here is nearest fit in some norm by Gauss.However it is
somewhat tricky to find the appropriate parameter values in (0,1) for
the intermediate controls.For more info refer to Hoschek/Lasser : Grundlagen
der geometrischen Datenverarbeitung.Those are the guys who wrote papers on the
parameter correction issue.Hoschek is my Thesis advisor right now at TH
Darmstadt Germany.Feel free to contact for more questions

---
  Markus Schichtel (joerg@cat.de)                     
  C.A.T. Kommunikations-System, Frankfurt