Skip to content

Latest commit

 

History

History
64 lines (60 loc) · 1.66 KB

vru_tracks.md

File metadata and controls

64 lines (60 loc) · 1.66 KB
Raw

Single vru track data format:

Class ids:

Class ID Class name Color
0 Pedestrian red
2 Bicycle blue
3 Motorcycle green
4 Scooter gold
5 Stroller pink
6 Wheelchair lime
10 Unknown gray

Track data:

  • ts : Timestamps are in UTC data format
  • status: Track classification status, 1 => Track Candidate, 2 => Accepted and valid track
  • class_prob: Probability of class identification in percantage 0-1
  • coordinates: Track 3d coordinates in intersection coordinate system in meters (m)
  • velocity: Velocity of object in kilometers per hour (kmh)
  • ground_type: Track ground classification
  • z_offset: Track height offset to intersection coordinate system origin
  • source_type: Which sensor(s) are responsible for current object detection, 0 => Lidar, 1 => lidar+camera, 2 => camera

Data format:

{
    "overview": 
    {
        "first_ts": val (int)
        "last_ts": val (int)
        "length": val (int)
        "duration": val (int)
        "class_id": val (int)
        "class_name": val (string)
    },
    "track_data": 
    {
        "ts_0": 
        {
            "ts": val (int)
            "status": val (int)
            "class_prob": val (int)
            "coordinates": list [x, y, z] (float)
            "velocity": val (float)
            "ground_type": val (int)
            "z_offset": val (float)
            "source_type": val (int)
        },
        "ts_1":
        {
            ...
        },
        "ts_2":
        {
            ...
        }
    }
}