Published:
After six months of work, the Path Matters project is submitted. This post is a summary of what we built, the results we got, and what I personally learned from it. The full report is available through TU Berlin.
Published:
This post documents the full 4-stage registration pipeline developed as part of the Path Matters project at TU Berlin. Starting from a VGGT reconstruction, we manually clean and align the point cloud using Open3D, then run BUFFER-X for coarse global registration, and finally refine with ICP — producing quantitative alignment metrics at every stage.
Published:
BUFFER-X Point Cloud Registration
Published:
An intelligent vision-based calibration system developed at Fraunhofer IPK and TU Berlin that reduces industrial robot TCP calibration time by 87.5% while improving accuracy by 76% through deep learning-driven pose selection.
