Point cloud labelling issues

[Odelllll]

Hi,
First of all, thank you very much for providing this conversion code.I have successfully converted the Pandar64 sensor data from the PandaSet dataset to the SemanticKITTI format and visualized the results correctly. However, when I attempt to convert the PandarGT sensor data, I encounter issues with the semantic labels during visualization. The labels appear to be incorrect or mismatched.
Here are the details of the problem:
I used the same conversion script that worked for Pandar64 data.
The point cloud data from PandarGT seems to be correctly extracted, but the corresponding semantic labels are not aligned correctly with the point cloud during visualization.
The labels appear to be incorrectly mapped or missing in some areas.
I suspect this might be due to differences in the data collection or the label mapping between Pandar64 and PandarGT, but I am unsure how to resolve it. Could you please provide guidance on how to properly convert the PandarGT sensor data, ensuring that the semantic labels are correctly aligned with the point cloud data?
Any insights would be greatly appreciated!

[SiMoM0]

Hi,
you should be able to extract the point cloud from PandaGT and the correspondening semantic labels, making this two modifications.

Set lidar sensor to 1 as scene.lidar.set_sensor(1) here:

Pandaset2Kitti/convert.py

Line 89 in c088af8

scene.lidar.set_sensor(0)

Then select the correct labels aslabels = scene.semseg[index].to_numpy()[-points.shape[0]:] here:

Pandaset2Kitti/convert.py

Line 101 in c088af8

labels = scene.semseg[index].to_numpy()[:points.shape[0]]

I just checked and it correctly display the corresponding labels to each point.

[Odelllll]

Thank you very much for your reply，I will try it!