Measurement System for AI-supported Detection of Surface Characteristics in Roll-to-roll Processes (SURFinpro)

The Fraunhofer Application Center for Optical Metrology and Surface Technologies AZOM, a branch of the Fraunhofer IWS in Dresden, has developed an intelligent measurement system (SURFinpro) for AI-supported detection of surface features in roll-to-roll processes (R2R). During the manufacturing of various layer systems or film systems based on R2R technologies, defects typically occur during the manufacturing process, which affect the outer appearance of the layers or the general quality and functionality of the systems. The structure of such manufacturing defects can manifest itself in a wide range of different sizes and characteristics. A laser triangulation approach is instrumentalized for the detection of the defects. Depending on the process, the components used can be adapted for optimal detection of the imperfections.

The actual core of the measurement system is the intelligent AI-supported evaluation of the data. Here, deterministic and statistical methods are used to measure the quality and properties such as roughness, scratches, defects or height distribution of surfaces. The processing of the image content for the extraction of the above-mentioned features is done using the realized inline scalable data infrastructure, which allows the execution of AI models as well as classical deterministic routines. Structures such as point defects as well as surface defects up to more than 500 mm can be reliably detected with a spatial resolution of about 100 μm at a speed of 5 m/min. The implemented auxiliary tools allow the flexible extension of the detectable feature set as well as the selection and evaluation of appropriate processing tools.

To the info sheet SURFinpro - AI-based Optical Surface and Defect Analysis for Continuous Production Processes

To the exhibition website SURFinpro