[feat] Aggregate point filter from a trajectory

Merged Julien Chaize requested to merge feat/addAggregatePointsFromTrajectory into master

Objective: Aggregate points from lidar data with a trajectory (containing transformation data).

  • Added a voxel grid based on an unordered_map to filter points. Added sampling options (first point found, last point found, max intensity, center point and centroid). The leaves of the voxel grid are uniform and described by a LeafSize. The bounding box of the voxel grid must be given as a parameter.

  • Added point aggregation filter. Sequential traversal of all the frames from the pcap. The transformation of each point is retrieved from the trajectory (via timestamp interpolation). The points are registered with this transformation and then filtered one by one via the voxel grid.

  • If requested, bounds can be automatically computed thanks to a sequential traversal of all the frames. The bounds of each frame are transformed using the trajectory and the bounding box of all the transformed bounds is calculated.


Corresponding slam documentation will be updated in keu-computervision/slam!253 (merged)

Edited by Julien Chaize

Merge request reports