This is a re-implementation of the 3Q3 solver (three quadratics in three unknowns) from Kukelova et al., Efficient Intersection of Three Quadrics and Applications in Computer Vision, CVPR 2016. The code is adapted from Jan Heller's original implementation. It also includes the additional tricks for improving stability based on choosing the elimination variable at runtime (see Zhou et al., A Stable Algebraic Camera Pose Estimation for Minimal Configurations of 2D/3D Point and Line Correspondences, ACCV 2018). Additionally we do a random affine change of variables to handle further degeneracies.
The function takes a 3x10 matrix corresponding to the coefficients of the three quadratics. The order of the monomials is
x^2, xy, xz, y^2, yz, z^2, x, y, z, 1.0;
There is also a mex-interface for matlab in re3q3_mex.cpp which can be compiled by running
mex('-I/usr/include/eigen3','re3q3_mex.cc','re3q3/re3q3.cc','re3q3/sturm8.cc')
For examples of how to use the library see PoseLib.
re3q3 is licensed under the BSD 3-Clause license. Please see License for details.