The categories of point cloud are confused.

[CallMeFrozenBanana]

CARLA version: 0.9.15 Platform/OS: ubuntu 20.04 Problem you have experienced: The class table in the example codes(open3d_lidar.py) doesn't match the output of semantic lidar sensor. What you expected to happen: I want to make clear the real label of the semantic lidar

This is the label in "open3d_lidar.py"

LABEL_COLORS = np.array([
    (255, 255, 255), # None
    (70, 70, 70),    # Building
    (100, 40, 40),   # Fences
    (55, 90, 80),    # Other
    (220, 20, 60),   # Pedestrian
    (153, 153, 153), # Pole
    (157, 234, 50),  # RoadLines
    (128, 64, 128),  # Road
    (244, 35, 232),  # Sidewalk
    (107, 142, 35),  # Vegetation
    (0, 0, 142),     # Vehicle
    (102, 102, 156), # Wall
    (220, 220, 0),   # TrafficSign
    (70, 130, 180),  # Sky
    (81, 0, 81),     # Ground
    (150, 100, 100), # Bridge
    (230, 150, 140), # RailTrack
    (180, 165, 180), # GuardRail
    (250, 170, 30),  # TrafficLight
    (110, 190, 160), # Static
    (170, 120, 50),  # Dynamic
    (45, 60, 150),   # Water
    (145, 170, 100), # Terrain
    (55, 90, 80), # Other 
]) / 255.0 # normalize each channel [0-1] since is what Open3D uses

Any suggestion is appreciated, thanks~!

[CallMeFrozenBanana]

Amazing that I found the correct class table here: https://carla.readthedocs.io/en/latest/ref_sensors/#semantic-segmentation-camera