Do the VLP-16 axes in Rviz match the axes in the User Manual?

[DavidB-CMU]

Do the cartesian axes of the PointCloud from the VLP-16 (as seen in Rviz) match the axes shown in Velodyne's User Manual?
For me it looks to be different.

http://velodynelidar.com/docs/manuals/VLP-16%20User%20Manual%20and%20Programming%20Guide%2063-9243%20Rev%20A.pdf

Here are the x,y,z axes specified in the User Manual:

This is those same axes in relation to the real sensor:

With the sensor in this same position, when I look at the PointCloud in Rviz, the axes of the /velodyne frame appear to be different by 1 rotation of 90 degrees.
You can see the x-axis and y-axis are in a different position:

[jack-oquin]

You are correct: this driver's coordinates do not match the diagram in the the VLP-16 manual.

The driver translates points into X, Y, Z coordinates the same way it does for other Velodyne models. It follows the standard ROS coordinate conventions from REP-0103 for axis orientation, so the X axis points to the "front" of the device.

Is that what you are seeing? Would you prefer it to work differently?

[DavidB-CMU]

Ok yes so I'm seeing the correct axes as per the ROS convention for cameras.

I don't mind, software is easy to modify, but it could be useful to add that to the documentation.
Thanks Jack. I can close this ticket.

[jack-oquin]

If you prefer to use the manufacturer's coordinate convention, it's easy to define a TF frame of reference that is 90 degrees to the right of the velodyne frame of reference.

[Kaiser1401]

I have a related question to this.

Is this 'Coordinate-System-Change' only done for the Velodyne 16? The axes from the 32 seem to match with the ones in the manual. I couldn't find a different handling of 16 and 32 though in the code yet.

[jack-oquin]

All models should work the same.