Where the curvature is not so pronounced XP may include line segments in the resulting polyline.
I agree that this is again related to short arcs and/or small radii but that statement won't hold here.
I can fool this a bit with a tighter approximation tolerance resulting in many more and shorter/flatter arc segments.
Even then I might get some line segments and thus corners as these are not tangentially connected.
Fewer line segments, that's true, but certainly a large number of short segments with negative consequences in a later stage.
The whole basic idea was a less tighter tolerance to avoid the so called 'small arcs'.
Example at hand (tolerance = 0.05):
- Line segment length: 0.09269668
Estimated arc segment:
Radius: 2.29026748
Sweep: 2.31915755°
Bulge: 0.01011958
Sagitta: 0.21833748 (9.5% of R)
- Adjacent arc segment chord length: 0.04545307
Radius: 2.04767242
Sweep: 1.27184527°
Bulge: 0.00554953
Sagitta: 0.24418725 (11.9% of R)
True, based on estimations the first listed segment was probably a fraction 'flatter'.
What is the determination method and threshold value for returning line segments instead of tangentially connected arcs?
Bottom line: Can it be forced to explode splines in polyline including nothing else then arc shapes?
Regards,
CVH