No Point Cloud Data Generated from RTX Lidar (Solid State) in Isaac Sim 4.5 Synthetic Data Recorder #3529
asivan-jpg
started this conversation in
General
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Hi,
I’m on Isaac Sim 4.5 (Windows 10) and trying to record point clouds from RTX Lidar (Solid-State) using the Synthetic Data Recorder. RGB images are saved correctly, but the point cloud .npy files are ~1 KB and effectively empty.
What I did:
Added RTX Lidar (Solid-State), positioned to look at the scene.
Enabled SD Recorder and turned on the LiDAR/Point Cloud channel (also enabled everything to be sure).
Pressed Play, then Record, waited a couple of seconds, stopped recording.
Files are created (e.g., pointcloud_0000.npy), but loading shows no data.
Loader snippet:
import numpy as np
arr = np.load("pointcloud_0000.npy", allow_pickle=True)
print(type(arr), getattr(arr, "shape", None), getattr(arr, "dtype", None))
If object array:
d = arr.item(); print(d.keys(), d.get("data", None) and d["data"].shape)
The files are ~1 KB; either an object array with no payload or (0,) sized arrays.
Tried/checked:
Timeline is running while recording.
Sensor pose/FOV aimed at near objects.
minRange < distances; maxRange > distances; resolutions > 0.
RTX renderer + RTX Lidar extension enabled.
No semantic filters that would drop all points.
Questions:
Is there any additional configuration required to record point clouds from RTX Lidar in Isaac Sim 4.5?
Do I need to enable a specific sensor output in the extension or in the recorder to make this work? which is the best camera or sensor to use in isaac sim in place with Tof sensors?
Has anyone successfully recorded point clouds using the RTX Lidar Solid State sensor with the Synthetic Data Recorder?
Beta Was this translation helpful? Give feedback.
All reactions