How to get the position of GCPs in the point cloud?

I did a 3D reconstruction of terrain using a drone and 8 GCPs. I would like, as validation, to measure the distance between the real measured position of GCP (that I give in gcp_list.txt) and its final coordinates in the point cloud.

I have been reviewing the code and, I think that in the point cloud the GCPs are not added. However, I have noticed that in opensfm/align.py there is a function triangulate_all_gcp that returns the triangulated and measured GCPs. I thought to use it and, for that, I wrote a small code where I read the undistorted reconstruction using data.load_undistorted_reconstruction() in order to pass it to triangulate_all_gcp function.

My question is: the final pose if it is in undistorted reconstruction?
I ask that because when I calculate the distance between both triangulated and measured GCPs, I get values of approximately 10 meters. Perhaps this is not the best way to do this.

Start with opensfm/align.py and the triangulate_all_gcp function, then inspect data.load_undistorted_reconstruction() and the point-cloud export path. Compare the triangulated and measured GCP coordinates with the final point cloud to determine whether they use the same coordinate frame; done means identifying the correct validation path or documenting the mismatch.