Hole-quality inspection using machine learning based on temperature images obtained using two-color high-speed video for Cu direct laser processes of a printed wiring board

This study aims to investigate a novel method for assessing processing quality by analyzing video data captured during substrate processing, instead of directly inspecting the substrate to be processed. Additionally, it seeks to investigate the relationship between processing phenomena and the shape of the processed holes. The advantage of this method is that inspections and machining can be simultaneously performed, leading to a reduction in both inspection time and the corresponding costs that would otherwise be required. Therefore, we first propose and evaluated an inspection method using multiple machine learning methods for multipulse machining. Next, we investigated whether the accuracy of anomaly detection varied based on the machined hole parameters.