Scanhead optical distortion correction

[Sortland]

Hi,

a rectangle drawn via a scanhead is typically distorted in a funny way (two sides bent to the inner, two to the outer side). This can be even stronger when some optics are used - and of course all geometries are affected, not only rectangles (but there one can see this easily).

My question: is there a formula available which can be used to correct these distortions? Until now I only know the possibility to have static (and not so accurate) correction tables...

Thanks!

[mixedgas]

A search for "Pincushion Distortion Galvanometer Scan Head" should bring you what you need.
*
Its predictable and correctable, or use a F-Theta lens..
*
It has everything to do with geometric optics and nothing to do with galvos, although in some cases you may need a slight tangent correction as well.

Steve

[VDX]

... for testing with XY-mirrors driven with servos instead of galvos and without a f-theta lens, I've simply measured the distortion and corrected it by software with a polynome.

As it's symmetric and radial 'regular', this was no problem with a quadratic equation ... for other calibration problems with much more nested variables I'm using Excels automatic calculated polynomics with up to 6 degrees or special programs and macros in Mupad (Matlab) ...

Viktor

[masterpj]

I usually correct by trial and error. Eventually you just get the feeling for it.
Alternatively if you don't have the capability to correct in software or the resolution of adjustement isn't high enough you can use an UGC.
I have one here and willing to sell it.

[Sortland]

@masterpj: I'm playing around with my own software, means I do not use any ready-to-use solution which is able to perform this correction. That's why I'm looking for a formula, I plan to do that for my own completely.

[White-Light]

There are plenty of formulas if you search the internet for pin cussion distortion, but the maths is complex.

The easiest I spotted was this which uses MatLab to do the work: http://uk.mathworks.com/matlabcentra...ion-correction

However, assuming there's no small matlab plugin type module you can / are allowed to include in your software, then:

This contains a formula for working it out yourself: http://www.vassg.hu/pdf/vass_gg_2003_lo.pdf

The same formula seem to be mirrored across many other sites, although there are plenty to choose from.

I'm kind of guessing you might need the latter if it's to include a calculation means in your software. Good luck, it looks complex.