Review of ORB-SLAM: a monocular SLAM system

ORB-SLAM

- Uses

  - Bundle Adjustment
  - ORB features [9]
  - A pose graph
     - Essential graph
     - a spanning tree
       - loop closure links
       - strong edges
       - from covisibility graph

  - covisibliity graph
    - local covisible area
    - tracking and mapping

- mar point and keyframe selection
  - generous spawning
  - restrictive culling
  - identify redundant keyframes
  - improves robustness and lifelong operations


- Stores map points:

  - 3D position X(w,i) in the world coordinate system
  - the viewing direction n(i)
    - the mean unit vector of all its viewing directions
    - the ray that joint the point with the optical center of the keyframe

 - A representative ORB descriptor D(i)
    - the associated ORB descriptor whose hamming distance is minimum
      with respect to all other associated descriptors in the keyframes
      in which the point is observed.

 - the maximum d(max)
 - the minimum d(min) distance
    - at which the point can be observed,
    - according to the scale invariance limits of the ORB features.


- Stores keyframes k(i)

  - The camera pose T (i,w)
   
    - a rigid body transformation
       - that transforms points from the world to the camera coordinate system

   - the camera intrinsics
      - including focal length and principal point

   - all the ORB features extracted in the framee
      - associated or not to a map point,
        whose coordinates are undistorted if a distortion model is provided.