As part of my algorithm, I need the arc length of a cubic function. The cubic function is generated by polyfit based off of my input data (in this case, the centroids of words in a line of skewed text).
For some (but not all) of my data, polyfit returns a warning stating: "Polynomial is badly conditioned." Based on the polyfit documentation I want to use the centering and scaling factors returned in MU to deal with this source of error. However, based on some of the methods used here, the best way to get arc length would be to differentiate my cubic function and then use the integral function.
Is there a way that I can easily rework the centering and scaling factors back into the polynomial coefficients so that I can pass them to polyder? Or will it be necessary to manually solve for the new coefficient values?
Best Answer