How to convert the k1 and k2 of your paper into the radial distortion parameters of the opencv checkerboard calibration method

[2585157341]

The minimization is performed in the normalized parameters p1 and p2 instead of k1 and k2. But I see the k1 and k2 are different from the k1 and k2 of the opencv checkerboard calibration method. How to convert the k1 and k2 of your paper into the radial distortion parameters of the opencv checkerboard calibration method? When I multiply the k1 by the square of the focal distance, k2 multiply the square of the focal distance, the undistion is still wrong.

[hugohadfield]

Hey, I also wondered this.. I had a look at the maths and have built a load of python bindings and parameter conversion functions etc. over here: https://github.com/hugohadfield/LensDistortionFromLines . Hope its helpful to you! I've added a bit of info of how to use them on the README.md

Also quick note, the parameter conversion I have done is into the opencv fisheye parameters, rather than the other opencv parameters. Although I imagine it won't be hard to convert to the normal radial ones, probably a bit easier than the fisheye ones as you wouldn't have to worry about the arctan(r) messing things up..