#33 Lens distortion effect on marker detection?

[Fnoop]

Camera intrinsic parameters are only applied to pose estimation, not to marker detection. Should images be corrected manually for lens distortion before passing to MarkerDetector, or can it cope with the distortion? I note in 'Speeded up detection of squared fiducial markers' conclusions it says: 'We consider as possible future works to investigate the use ofthe proposed method in fisheye cameras, as well as to characterizethe performance when multiple fiducial markers with significantlydifferent scales are present in the same image'.

I am also using a specific set of markers as a concentric landing pattern for UAVs:
https://github.com/goodrobots/vision_landing/blob/master/markers/april16h5-landing-a1.png
As I understand from the new paper the system calculates the adaptive to the smallest marker found? Is there a way of stopping the detection calculations after finding the smallest marker? I'm interesting in the fastest possible method of finding the smallest marker - ie. I'm only interested in the smallest marker at any one time.

Thanks for your great work on the new code, very interesting paper.