Class ID Mismatch in ros2_camera_helper: Only 13 Classes Detected Despite 23 Annotations

Omniverse Isaac Sim
,
1

I used ros2_camera_helper to get bbox3d data, but the class_id only includes 13 classes. I annotated 23 characters (e.g., character_0, character_1, etc.) using the annotation tool, yet it still only recognizes 13 classes.

2

Moving this to the IsaacSim forums where they can assist you better.

3

Could you help clarify what “ros2_camera_helper” refers to? Also, could you provide detailed steps to reproduce this issue?

4

graph coding, ros2_camera_helper node, which can publish topic like img or bbox3d in img, and I connect it with a camera, using it to publish bbox3d messages, I annotate 23 human agent objects with 23 types, like character_0, character_1, But, It comes out only 13 class_id from

  • header:
    stamp:
    sec: 0
    nanosec: 0
    frame_id: ‘’
    results:
    • hypothesis:
      class_id: ‘6’
      score: 1.0
      pose:
      pose:
      position:
      x: 0.0
      y: 0.0
      z: 0.0
      orientation:
      x: 0.0
      y: 0.0
      z: 0.0
      w: 1.0
      covariance:
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
      • 0.0
        bbox:
        center:
        position:
        x: 1.5493332147598267
        y: -14.507498741149902
        z: 0.02103900909423828
        orientation:
        x: 0.0
        y: 0.0
        z: 0.0008365732858670433
        w: 0.9999996500725075
        size:
        x: 0.4668655529978795
        y: 0.3321409321130205
        z: 1.6910808086395264
        id: ‘’
5

will have class_id = 6 for not only one

6

Could you share detailed reproduction steps for this issue to aid our investigation?

7

yes, there are reproduction steps below

① annotate like this

image
image879×859 114 KB

each person have a unique semantic type

② use omnigraph graph’s ros2 camera helper to publish bbox3d

image
image241×330 39.5 KB

③ eventually, we gathered bbox3d message, but not each person have unique class_id, some person have the same class_id when annotated different semantic type.