Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Point cloud to world frame transformation #134

Closed
m-decoster opened this issue Feb 29, 2024 · 3 comments · Fixed by #141
Closed

Point cloud to world frame transformation #134

m-decoster opened this issue Feb 29, 2024 · 3 comments · Fixed by #141
Labels
enhancement New feature or request

Comments

@m-decoster
Copy link
Contributor

Describe the feature you'd like
When working with point clouds, sometimes we want to express them in the world (or robot base) frame.
I currently do this with a method proposed by @Victorlouisdg through Open3D:

pcd_in_camera = point_cloud_to_open3d(point_cloud)
pcd = pcd_in_camera.transform(X_W_C)
point_cloud = open3d_to_point_cloud(pcd)

Do we want to create a function for this in the camera toolkit (leaving possibilities for future optimisations, I haven't checked the performance implications yet of the above conversions to and from Open3D. I'm assuming it's fairly cheap because it's zero copy)?

Use cases

  • Merging point clouds from different cameras (e.g., with ICP)
  • Logging point clouds to, e.g., rerun

Possible implementation
We could wrap the above code snippet into a function named transform_point_cloud, which could go into https://github.com/airo-ugent/airo-mono/blob/main/airo-camera-toolkit/airo_camera_toolkit/point_clouds/operations.py
@m-decoster m-decoster added the enhancement New feature or request label Feb 29, 2024
@tlpss
Copy link
Contributor

tlpss commented Feb 29, 2024

Actually you can do the following w/o going to open3D:

# assumes pc is an Nx4 array with homogeneous points. 
pc_in_A = (X_A_B @ pc_in_B.T).T

In terms of adding a function for this, we could do that if required/desired.

I think the pc operations are indeed the best place to add them. There is actually already a helper class in the spatial-algebra package, that could be used, we can simply wrap it: https://github.com/airo-ugent/airo-mono/blob/main/airo-spatial-algebra/airo_spatial_algebra/operations.py.

@m-decoster
Copy link
Contributor Author

So if I get it correctly, we would only need to wrap operations.transform_points, because the point cloud positions is a Vectors3DType.

Since Open3D conversion is not required in that case, then perhaps adding an additional function for this is overkill.

(For me this was about reducing a 3-line operation + imports to a one-liner, but transform_points does the same)

@tlpss
Copy link
Contributor

tlpss commented Feb 29, 2024

So if I get it correctly, we would only need to wrap operations.transform_points, because the point cloud positions is a Vectors3DType.

indeed!

Since Open3D conversion is not required in that case, then perhaps adding an additional function for this is overkill.

on the other hand it will be easier to find. So I think it could still be a good idea

@m-decoster m-decoster mentioned this issue Mar 7, 2024
Merged
2 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
enhancement New feature or request
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Point cloud transform airo-ugent/airo-mono
3 participants
@m-decoster @Victorlouisdg @tlpss