Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.


dpb
Posts:
9,534
Registered:
6/7/07


Re: curve interpolation by parabolic blending
Posted:
Jun 22, 2013 7:08 PM


On 6/22/2013 7:06 AM, enprush@gmail.com wrote: > I'm given a set of 25 points(x,y values). I need to curve fit these > points using a parabolic blending. I'm not sure how to go about it. I > read that we need to take 3 points at a time to ensure a unique parabola > is obtained. > I need to: > take first 3 points(1,2,3) and fit a parabola to it. > take the next 3 points (2,3,4) and do the same. > This goes on till i reach my final point. The overlapping portions > of any two parabolas is linearly blend using a cubic curve. > > So, this looks more like a polyfit of third order except between > point 1 and 2. Am i right in continuing on with a ployfit of order 3?
Well, you can certainly find the coefficients of the quadratic at each set of three points w/ polyfit and use polyval to evaluate it over the range.
I don't know the details of how the blending is supposed to work so can't suggest much on that point w/o having some reference to work from. A quick google turned up several references but none that weren't more reading that had time to delve into...




