Skip to content

jhultman/continuous-fusion

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

46 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Continuous fusion

C++ implementation (in ROS) of continuous fusion algorithm from Deep Continuous Fusion for Multi-Sensor 3D Object Detection. I can share a fast, vectorized SciPy version if there is any interest (update: see issues page).

Fusion

Comments on coordinate systems:

Points x_velo in velodyne coords are sent to points x_image
in the image plane of camera 2 using (7) in Geiger et al.:

    x_image = P * R * T * x_velo,

where:

    x_velo  - point in velodyne coords (x, y, z, 1).
    T       - velodyne coords to unrectified cam0 coords.
    R       - unrectified cam0 coords to rectified cam 0 coords,
    P       - rectified cam0 coords to image plane of cam2,
    x_image - point in image of cam2 (u, v, 1).

(We use left-multiplying active/alibi coordinate transformations.)

Diagram borrowed from paper (we skip MLP):

Fusion

Releases

No releases published

Packages

No packages published

Languages