Making a checkerboard for UV Camera Calibration

I’m currently trying to perform camera calibration on a camera lens with a UV filter on it using the checkerboard method. I’ve had a very tough time making the checkerboard visible with enough contrast for the FindChessboard algorithm to work successfully. I’ve been using a laser printed chessboard that I’ve been hanging outside. With the right camera settings and enough sunlight, I can get the image to appear with enough contrast for the checkboard algorithm to work. See below:

However, the success of this method is inconsistent, and I need it work consistently inside as well as outside.

Does anyone have any ideas for how I could make this work better and more consistently? Should I still be trying to use the chessboard method, but use different materials for the chessboard (if there’s anything that can reflect UV or be picked up easily on a UV filtered lens)? Are there other methods for undistorting a lens that would work better for a UV filtered lens?

Thanks for any direction or help.

1 Like

welcome.

picture looks noisy so I’d say it’s a subset of those:

  • camera contains an UV-block filter (also commonly blocks IR)
  • UV illumination is weak
  • your UV-pass (everything-else-block) filter is passing hardly any UV
  • camera aperture and exposure time are insufficient

remedies derive from the opposites of each.

  • remove UV-block filter from camera assembly
  • moar UV: wear goggles if you value your corneæ
  • different filter? perhaps test this one using photo film and plain UV light, compare exposure times with vs without your UV-pass filter
  • longer exposure time, perhaps “image stacking”/exposure stacking, i.e. sum exposures for an artificially longer one than the camera lets you do normally

the checkerboard should be the best choice here. a circles grid will become inaccurate when tilted out of plane (but it’s convenient sometimes), and more advanced patterns such as “charuco” require less noise/better resolution than you have there.