Guard improves robotic part detection by filtering uncertain 3d points
GUARD: Geometric Uncertainty-Aware Point Cloud Denoising and Segmentation for Robotic Hard Disk Drive Disassembly
Computer Vision and Pattern RecognitionRobotics
Summary
Robots that take apart hard disk drives need to tell real parts from scanning errors in 3D images. The authors created GUARD, a method that can spot which parts of a 3D scan are reliable and which are likely mistakes. GUARD uses smart math to check uncertainty in the scan shapes and cleans out errors while keeping important details. Tests show GUARD helps robots more accurately identify parts in complex 3D data, especially when scans have noise or mistakes.
What this means in practice
- •For robotic system builders: Enhance robotic disassembly systems for devices like hard disk drives by reliably distinguishing true parts from scanning artifacts in 3D point clouds.
- •For 3d vision engineers: Improve point cloud segmentation algorithms by incorporating geometric uncertainty to detect and suppress corrupted data while preserving useful structure.
Authors
Zuoxu Wang, Xiao Liang
Abstract
Reliable robotic disassembly requires part-level representations that distinguish genuine component geometry from scanning and reconstruction artifacts. In point clouds of hard disk drives (HDDs), structured ghost artifacts can resemble valid components locally while remaining inconsistent with the overall geometry, allowing erroneous measurements to receive plausible semantic labels. This creates an engineering information problem: semantic prediction confidence alone does not establish whether the underlying geometry is reliable. We propose \textbf{GUARD}, a geometric uncertainty-aware framework that performs point filtering and segmentation within a single forward pass by modeling the reliability of learned geometric representations. GUARD combines a multi-scale geometric transformer with a multi-bandwidth random Fourier feature Gaussian Process to estimate per-point geometric uncertainty, complemented by predictive entropy to suppress unreliable measurements while preserving informative structures. Evaluation on 2,745 real HDD point clouds shows that GUARD improves PointNet++ segmentation mean intersection over union from 0.7739 to 0.8318. Additional experiments on ShapeNetPart and ScanNet examine robustness across corruption types, point-cloud domains, and segmentation backbones. On manually annotated ScanNet samples, geometric uncertainty achieves a corrupted-point detection F1 score of 0.7931, compared with 0.2212 for predictive entropy. The results demonstrate the value of distinguishing geometric reliability from semantic confidence and reveal a tradeoff between artifact suppression and preservation of informative structures. GUARD contributes a reliability-aware approach to interpreting imperfect 3D measurements for component identification and subsequent robotic handling. Project website: https://001-wang.github.io/GUARD_Point_denoiser/.