In ZED fu, we extracted the fused point cloud from the saved SVO file.
We did several times of the extraction from the same SVO file, but the extracted point cloud sets are different every extraction.
Please let us know why it happens.
And please give us some comments to extract good point clouds from a SVO file. Is there a specific setup or some steps for the extraction in ZED fu?
If the depth_stabilization InitParameter is set to 0, and the depth used is not NEURAL, it should be deterministic.
It seems that in ZEDfu, you can’t set it to 0 however, I’ll log this for fixing. That being said, the results should be very similar from one run to another.
- Are there notable differences in the extracted point cloud sets?
Getting a good point cloud from ZEDfu involves the same steps as getting a good point cloud from a real-time application:
- good lighting
- non-uniform surfaces will give the best results
I hope this helps