onavhamid / veloparser Goto Github PK

Veloparser is a simple application which does following:

• Supports just Velodyne VLP16 at the moment.
• Takes a pcap file recorded by Velodyne lidar as input.
• Extracts all Frames from the pcap file.
• Saves both data-frames and position-frames.
• Data frames are saved as Point Clouds (.pcd) and/or as plain Text-File.
• Position frames are saved only as Text-File
• Converts frame's timestamps to GPS Week of Second format for synchronization with IMU/GNSS devices
• Can be parameterizes by yaml file.

The reason why i wrote it, is simply that i could not find any simple way without installing ROS (Robot operating software) or other huge c++-based lib that does 'just' extract the point clouds and GPS-Timestamps from pcap-file.

Assuming using Anaconda python distribution:

~$ cd \path-to-veloparser-repo\
~/veloparser$ conda env create -f environment.yml
~/veloparser$ conda activate veloparser
~/veloparser$ python main.py -p /home/user/my.pcap -o /home/user/output_folder -c params.yaml

Note, the params.yaml can be updated according to your setup. For example, specifying whether GPS was available.

Veloparser has following package dependencies:

• dpkt
• numpy
• tqdm
• yaml

Please make sure that all of those packages are installed (pip or conda).

Below a sample out of 2 Points in a point cloud file

Time [musec], X [m], Y [m], Z [m], ID, Intensity, Latitude [Deg], Longitudes [Deg], Distance [m] 2795827803, 0.032293, 5.781942, -1.549291, 0, 6, 0.320, -15.000, 5.986 2795827806, 0.083565, 14.399564, 0.251350, 1, 6, 0.333, 1.000, 14.402

All Point Cloud PCD-Files have following fields:

1. X-Coordinate
2. Y-Coordinate
3. Z-Coordinate
4. Intensity

They can also be opened and visualized with any point-cloud rendering software like (open3d, pcl, ...)