British Machine Vision Conference (BMVC), 2015

Ekaterina Kanaeva, Lev Gurevich, Alexander Vakhitov

We propose a new method for camera pose estimation with unknown focal length (PnPf problem). We combine projection equations and distance constraints in a single statistically meaningful cost function in the form of least squares. We fix the space of the search as a linear combination of several right singular vectors of the least squares system matrix. We use linear programming techniques to find feasible solutions faster. Then we do nonlinear refinement with Levenberg-Marquardt. Numerical experiments demonstrate that the method is faster than the state-of-the-art methods for point numbers up to several hundreds, and real-life structure-from-motion experiments demonstrate the applicability of the methods for models having hundreds of thousands of points. It has the same accuracy of estimates as the the state-of-the-art methods. We show that the method offers a tradeoff between speed and accuracy, allowing the estimation to run several times faster while slightly increasing the mean reprojection error.

Code

MATLAB implementation

Paper

Kanaeva, Ekaterina and Gurevich, Lev and Vakhitov, Alexander Camera Pose and Focal Length Estimation Using Regularized Distance Constraints, BMVC, 2015, pp. 162–1 pdf bib