Circle fitting using Kåsa's method
Example:
# import library
using CircleFit
using StatsBase
# generate test data
r = 5
x0 = 2
y0 = 4.5
x = r.*[-1.0,0,0,1] .+ x0
y = r.*[0.0,1,-1,0] .+ y0
# fit
result = fit(Circle,x,y)
# fitted coefficients
coef(result)
# (2.0, 4.499999999999999, 5.0)
# names of the coefficients
coefnames(result)
# ("center position x1", "center position x2", "radius")
# calculate the residuals square sum
rss(result)
# 1.5777218104420236e-30Gradient weighted algebraic fit for a circle using LsqFit.levenberg_marquardt: CircleFit.GRAF(x,y,p0) or fit(Circle,x,y,alg=:graf)
Non optimized implementations:
- Method by Taubin
CircleFit.taubin(x,y)orfit(Circle,x,y,alg=:taubin) - Method by Pratt
CircleFit.pratt(x,y)orfit(Circle,x,y,alg=:pratt)