Having dramatically different step sizes, and not much data. We have data like >--4-6----3-9----4-7----9-3----4-1----> ( - is step here ) and in real life this data is looped in a circle. What polynomials/formulas can help me with interpolation of such data? Will Bezier work?
Asked
Active
Viewed 96 times
0
Brian Tompsett - 汤莱恩
- 5,753
- 72
- 57
- 129
Rella
- 65,003
- 109
- 363
- 636
-
Fitting a smooth curve to discrete data is often misleading. – job Dec 14 '10 at 21:48
-
When you say "looped in a circle" do you mean the data is periodic? Do you know the period? – mtrw Dec 14 '10 at 22:01
-
Does the interpolated curve have to pass through the control (i.e., given data) points? – ysap Dec 14 '10 at 23:19
1 Answers
1
Fit Cublic Splines on each set. A single BeZier curve is not accurate enough (unless you know otherwise). Another option is Chebyschev Polynomials.
John Alexiou
- 28,472
- 11
- 77
- 133
-
IIRC, Beziers don't pass through the control points, so I am not sure it will fit the OP's requirement. OTOH, the OP did not really mention if this IS a requirement. – ysap Dec 14 '10 at 23:18