Skip to content

Latest commit

 

History

History
19 lines (17 loc) · 1.24 KB

File metadata and controls

19 lines (17 loc) · 1.24 KB

Reconstructor

Camera position reconstructor for Blender, using correlated motion tracker markers across different movie clips to solve camera positions

General Plan/Workflow

  1. Create motion tracks from clips with matching focal lengths.
    • Individual images within clips must have the same focal length
    • Clips should be at least partially track-solved (if correspondences found) to derive optical center or distortion params.
  2. ADDON: Create cameras for each clip.
    • Alternately, can link clips to existing cameras.
    • Define rough axial position wrt target object (i.e. front, above, left) to prime/constrain solve
    • Crux: cameras correlate with clips.
  3. ADDON: Create empties for each tracked element.
  4. ADDON: Associate correlated empties from different clips, so that their "solves" can converge
  5. ADDON: Solve -> Bundle adjust linked empties and cameras to minimize distance between linked empties
    • Beware converging the cameras toward 0
  6. RESULT: Cameras and empties positioned in best approximation of analogous 3D positions, with minimized error

For mathematical details underpinning the concept, see MATH.md.