better way to calibrate

[thereza]

Rather than using a ruler, I found that printing out some mm based grid paper and using that works better. Ideally, the software would be able to see the lines, and compute a mapping to be able to algorithmically correct optical nonlinearities.

[JuKu]

Indeed, this is a much better idea!

> Ideally, the software would be able to see the lines, and compute a mapping to be able to algorithmically correct optical nonlinearities.

Yes, ideally. On the other hand, at least this example does not show much error on the center of the image, which is what matters.