Skip to content

Latest commit

 

History

History
35 lines (29 loc) · 1.5 KB

robust_6dof_motion_multicam.md

File metadata and controls

35 lines (29 loc) · 1.5 KB

Robust 6DOF Motion Estimation for Non-Overlapping, Multi-Camera Systems

Author: Clipp

Year: 2008

Notes:

  • pose estimation with scale with non overlapping FoV setups
  • assumption: 5 temporal correspondences in cam + 1 corespondence in any additional camera
  • Derive the expression of the scale using a single point correspondence $x \leftrightarrow x'$ : $$ \lambda = -\frac{x'^TAx}{x'^TBx} $$
  • Geometrical explanation: the view of a 3D point seen from the first camera crosses a plane made by the different $C_2(\lambda)$ and the ray of $x'$ noted $v'$. The $v'$ that points to this intersection point fixes the scale
  • Critical configurations:
    • $v$, $v'$ and $C_2'(\lambda)$ can't be coplanar
    • Far away points ($\pi$ and $v$ are parrallel)
    • The axis of rotation is perpendicular to the plane made by $C_2, C_1'$ and $C_1$
  • Check for critical configurations: double the scale and check if there is still a big fraction of inliers

Algo:

  • 5pt algorithm + RANSAC + Chirality check to estimate motion up to scale
  • One point RANSAC + pbM estimator to recover scale
  • Inlier correspondences on the second camera are determined using the distance to epipolar lines
  • then a linear least square calculation of the scale factor is made with inlier correspondences
  • Then a bundle adjustement of the full multi camera system is performed

Exp:

  • simulations experiments with randomly generated points in a cube
  • real experiments using 8 cameras
  • Only the scale degrades when the motion becomes ambiguous