In the last decades, machine vision and Machine Learning (ML) techniques have seen significant improvements in the development of new algorithms thanks to the increment of hardware performance. For this reason, applying computer vision for specific technological problems became an important opportunity to introduce some significant improvements in the manufacturing context. Indeed, several studies on the application of ML in the manufacturing process are available. A good application of ML is monitoring the qualitative aspects of a manufacturing process. This paper proposes a preliminary study to analyze the ML capabilities to perform Automated Optical Inspection (AOI) for quality control in the manufacturing of Printed Circuit Boards (PCBs). In this specific case, the target has been to investigate the performance of the method Mask R-CNN to individuate the main PCB defects after the manufacturing process. This study has been performed considering an available open-source dataset employed by other ML techniques.
For this reason, this study has aimed to verify the effectiveness of the adopted ML solution to manage this application. The chosen open-source dataset individuates the opportune class of products and related defects for the context of interest. In this specific case, this work has been carried out to gather know-how for further activity related to AOI for quality control in the assembly of PCBs employed in the aerospace field.