Advanced manufacturing requires both large-area process monitoring and high-precision local measurements under complex and high-speed production conditions. In semiconductor, display, and battery manufacturing, small variations in thickness, surface profile, and defect distribution directly affect device performance, yield, and long-term reliability. As production speeds increase and device structures become more complex, metrology systems must provide not only high spatial resolution but also stable and repeatable performance during in-line operation. To address these multi-scale challenges, we present an optical metrology framework that combines AI-based vision for wide-area inspection with spectral-domain interferometry for point-based nanometer-scale measurement. The framework is demonstrated through laser-based cutting of large glass substrates used in display manufacturing, where high-speed in-line processing can generate micro-cracks as well as local thickness and refractive index variations caused by thermal and mechanical effects. The AI vision system enables fast crack detection and spatial analysis over large moving substrates without interrupting production, while spectral-domain interferometry provides non-contact thickness and refractive index measurement at selected points for detailed analysis. Although the two systems operate independently according to process needs, their architectures are developed under an international traceability system based on the Guide to the Expression of Uncertainty in Measurement (GUM) to ensure reliable and consistent measurement results across different production runs and sites. By combining large-area inspection and precise point-based measurement within a unified reliability-focused framework, the proposed approach provides a practical solution applicable not only to glass processing but also to semiconductor wafer and battery manufacturing.
