From: Hans-Bernhard B. <HBB...@t-...> - 2017-02-20 21:57:58
|
Am 20.02.2017 um 21:45 schrieb Peter Williams: > It looks like the routine is failing to join successive > points of the closed contour together. The routine cannot join points in positions if there aren't any in the place you think they should be. In particular, you do not have any samples at y=0, because a) your 'isosamples' setting is even (should be odd) > set isosamples 256,256 b) you didn't "set samples", so it also defaults to even 100, but you want it odd Generally speaking, your expectation that a discrete contouring algorithm will _exactly_ locate a point-like feature like this saddle point is somewhat ill-founded. |