AI Transformation in Optical Coating Inspection

Technical Case Study

This case is designed to help industry professionals understand the R&D capabilities of the Fraunhofer Innovation Platform for Surface and Production Engineering for Optical and Electrical Systems at Feng Chia University (FIP-SPE@FCU). Using a recent technical paper on "AI-based Automatic Detection for Optical Coatings" as a case study, we demonstrate how we solve industrial quality control challenges and outline our comprehensive service portfolio.

Microscopic Thin-Film Defect Detection

Core Technology

In the mass production of optical thin films (e.g., filters, lenses), microscopic defects like scratches, stains, or bubbles are inevitable. Traditional manual microscopic inspection is time-consuming, and operator fatigue leads to missed detections, making it difficult to meet modern demands for high throughput and high yield.

The core of this technical solution is the integration of high-precision image acquisition with deep learning technology to achieve a fully automated solution for microscopic defect detection on optical thin-film surfaces.

Standard System Operational Workflow

The operation of this automated inspection system on the actual production line is continuous and real-time. The workflow is as follows:

Before vs. After Implementation

The following comparison highlights the significant differences between traditional manual inspection and the AI detection system developed.

The results of the AI detection identifying original thin-film images of various microscopic thin-film defects.