PoseEstimation

Overview

• C++ template library for:

  • EPnP algorithm [1,2] for projective-n-point
  • P3P algorithm [3] for projective-3-point
  • Five point algorithm [4] for Essential matrix esitimation
  • Seven point algorithm for Fundamental matrix esitimation (almost equivalent to [5])
  • Three point algorithm [6] for 3D similitude transformation estimation

• These algorithms support:

  • RANSAC scheme for robust estimation
  • double / float precision as input
  • unit test by using CPPUnit (except for EPnP)

Usage

Dependencies

• CMake
• Eigen 3
• CPPUnit (optional)

Build

$ cd PoseEstimation
$ mkdir build
$ cd build
$ cmake -DCMAKE_VERBOSE_MAKEFILE=true -DCMAKE_BUILD_TYPE=Release ../src
$ make -j8
$ make test

References

[1] V. Lepetit, F. Moreno-Noguer and P. Fua. "EPnP: An Accurate O(n) Solution to the PnP Problem". International Journal of Computer Vision, vol. 81, pp. 155-166, 2009.

[2] F. Moreno-Noguer, V. Lepetit and P. Fua. "Accurate Non-Iterative O(n) Solution to the PnP Problem". In Proc. of IEEE International Conference on Computer Vision, Rio de Janeiro, Brazil, October 2007, pp. 1-8.

[3] L. Kneip, D. Scaramuzza and R. Siegwart. "A novel parametrization of the perspective-three-point problem for a direct computation of absolute camera position and orientation". In Proc. of IEEE International Conference on Computer Vision and Pattern Recognition, Colorado Springs, CO, USA, June 2011, pp. 2969-2976.

[4] D. Nistér. "An Efficient Solution to the Five-Point Relative Pose Problem". IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 26, no. 6, pp. 756-777, 2004.

[5] O. Faugeras. "Stratification of three-dimensional vision: projective, affine, and metric representations", Journal of the Optical Society of America A, vol. 12, issue 3, pp. 465-484, 1995.

[6] B. K. P. Horn. "Closed-form solution of absolute orientation using unit quaternions", Journal of the Optical Society of America A, vol. 4, issue 4, pp. 629-642, 1987.