On-machine dimensional inspection: machine vision-based approach

The contemporary industry has witnessed a significant transformative development with the integration of artificial intelligence (AI) in various industrial systems, resulting in an enhanced automation for heightened productivity and efficiency. However, mastering this level of automation can be challenging for some applications, such as manufacturing inspection, which can be delicate while maintaining a precise cadence for an in-line manufacturing scale. In this paper, a systematic machine vision-based approach for on-machine inspection is proposed in order to automate and improve inspection process towards computer numerical control (CNC) machined parts. The approach incorporates remapping algorithm and image processing operations to accurately extract desired features. Subsequently, these features will undergo dimensional inspection based on their generated point clouds. Tests were applied on a sample part using a complementary metal–oxide–semiconductor (CMOS) camera mounted on the spindle of 5-axis CNC machining center. The paper explores numerous aspects related to different stages of the approach and their impact on the resulting inspected features evaluations. It also highlights significant findings regarding critical factors for conducting well-structured experiments at various stages. Promising results have shown the significance of the presented work regarding industrial automation technology, ultimately improving manufacturing efficiency throughout the production line.