Automated 3D burr detection in cast manufacturing using sparse convolutional neural networks

Abstract

For automating deburring of cast parts, this paper proposes a general method for estimating burr height using 3D vision sensor that is robust to missing data in the scans and sensor noise. Specifically, we present a novel data-driven method that learns features that can be used to align clean CAD models from a workpiece database to the noisy and incomplete geometry of a RGBD scan. Using the learned features with Random sample consensus (RANSAC) for CAD to scan registration, learned features improve registration result as compared to traditional approaches by (translation error ($$\Delta $$ Δ 18.47 mm) and rotation error($$\Delta 43 ^\circ $$ Δ 43 ∘ )) and accuracy(35%) respectively. Furthermore, a 3D-vision based automatic burr detection and height estimation technique is presented. The estimated burr heights were verified and compared with measurements from a high resolution industrial CT scanning machine. Together with registration, our burr height estimation approach is able to estimate burr height similar to high resolution CT scans with Z-statistic value ($$z=0.279$$ z = 0.279 ).

3D Pointcloud Registration In-the-wild

“…Registration of cast manufactured parts is a prerequisite for automation of tasks such as sanding, welding and assembly in low-volume production. Typically, the part is placed with an arbitrary pose on a table surface or bin, and scanned 1) Cast manufactured parts dataset: We use 3D data of cast manufactured parts acquired for the scan-to-cad dataset in [16] to construct a scan-to-scan dataset. The cast parts are brass parts from from Mjøs Metallvarefabrikk, approximately 30-40 cm in diameter, and 20 individual physical items.…”