OpenGV is a collection of computer vision methods for solving geometric vision
problems. It contains absolute-pose, relative-pose, triangulation, and
point-cloud alignment methods for the calibrated case. All problems can be
solved with central or non-central cameras, and embedded into a random sample
consensus or nonlinear optimization context. Matlab and Python interfaces are
implemented as well. The link to the above pages also shows links to precompiled
Matlab mex-libraries.  Please consult the documentation for more information.