Yes, there is no direct conversion possible unless you limit the parameters. However given the two formulas, this is a curve fitting problem - it ought to be possible to get a very good approximation with a spreadsheet solver.
-- Bruno On 1 July 2016 at 08:25, Torsten Bronger wrote: > Hallöchen! > > Finfa811 writes: > >> I've found a solution in the following link: >> >> *http://www.panotools.org/dersch/barrel/barrel.html* >> >> - a = k2 * (width/2)3 >> - b = 0 >> - c = k1* (width/2) >> - d = 1 >> >> The minimum input that OpenCV uses for its polynomial is k1k2p1p2 >> (4 parameters). In my case, just the parameter c was not zero >> (other than d), so only k1 will be used. The result is pretty much >> the same. > > I'm suprised that this works. > > First, the only coefficients of the same power in r in both models > are b and k1. Both apply to r^3. OpenCV's k2 and k3 apply to r^5 > and r^7, respectively, both of which don't have counterparts in > Hugin. > > Secondly, d in Hugin is 1-a-b-c rather than 1. The difference is > noticable for real-world lenses in my experience. > > I think the only way to make both compatible is to use only b in > Hugin for calibration, take the different coordinate system into > account (i.e. scaling factor to the power of 3), and take d!=1 into > account by dividing the resulting k1 by (1-b)^3. > > (Beware that I haven't checked the calculation, and one can easily > make mistakes in these things.) -- A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ --- You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/CAJV99ZhbPnHBq2pkBmmeWA1Ctkw9UVNKdfnTwvUvz-WwBtkZ%2BA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
