45
There are two versions of point cloud files that need to be converted. The earliest version from
Stencils with software up to February 2018 release where the full timestamp is embedded with
each x, y, z point. The July and November 2018 releases included an offset timestamp in the
header with the deltas from the offset included with each x, y, z point. Both of these approaches
caused issues when working with 3
rd
party software. Kaarta moved these key times and position
offsets (from GNSS) to a machine readable .yaml file so these values could be shared between the
operating modes of Stencil, features, and the post-processing tools like loop closure. Users can
also access these key values using the scripting support tools described below. It also allows
editing of the point cloud, trajectory, and GPS .ply files by 3
rd
party software without impacting the
other post-processing tools.
Modify timestamp of every point in a ply file
– This utility provides an easy way to change the
time offsets in any of the Kaarta .ply files.
rosrun kaarta_io modify_ply_time [filename input] [filename output] [time offset]
Down sampling large point clouds by time
– This utility will down sample a point cloud by time
to the number of points in the rosparam set command.
rosparam set /velodyneCloudGenerator/maxPointNum 60000000
roslaunch clay_launch generate_map.launch in_file_name:=<path to input file with
name> out_file_name:=<path to output file with name> useDefaultPointNum:=true
Remove moving objects from the point cloud
– This utility provides an easy way to eliminate
points associated with moving objects such as people or vehicles. After running, be sure to look at
the point cloud to make sure you did not remove too much of the sparser structural components.
rosrun pointcloud_cleaning movingObjectFilter _pointcloud_in:= [pointcloud input
filename] _pointcloud_out:=[pointcloud output filename] _trajectory_in:=[trajectory input
filename] [optional parameters]
