Include descan error and other distortions

[uellue]

At https://er-c-data-mgmt.fz-juelich.de/browse/?path=/adhoc/livecalibration/thermo-empad/raw/rotation_calibration are some datasets that show strong descan error and also distortion of the shape of the aperture as a function of scan position.

[uellue]

The shift component of a point in scan coordinates introduced by descan error should be independent of the overfocus, while the component introduced by scanning is anti-proportional to the overfocus (small overfocus -- large magnification -- large shift as a function of scan). With two or more measurements at different overfocus it should be possible to distinguish the two influences.