Surface Monitoring for Additive Manufacturing with in-situ Point Cloud Processing

Abstract

Vision-based surface monitoring is critical for surface defects and geometric distortion detection during additive manufacturing (AM) processes. It is the pre-requisite of quality assurance and process control to maintain and improve the quality of AM produced parts. However, current surface monitoring solutions are not efficient enough to provide in-situ monitoring feedback during the relatively long AM fabrication process, which has become a significant barrier to AM automation. This paper presents a new surface monitoring method integrated in a robot-based laser-aided additive manufacturing cell, with in-situ point cloud processing capability. During surface scanning, a multiprocessing technique is used to run both sensor data capturing and point cloud processing programs. Point cloud data captured within predefined time intervals are filtered and segmented automatically to extract the parts surface to be monitored. Experimental results are presented to verify the effectiveness of the proposed surface monitoring method.