Abstract
This study presents an automated system for visual inspection of optically challenging composite surfaces with highly specular binders and weave materials, where conventional inspection methods fail and manual inspection is still common. Our approach—linear lighting combined with off‑axis imaging—captures subtle (<0.1 mm) surface defects by detecting disturbances in the bidirectional reflectance distribution function (BRDF). The underlying imaging physics and analysis methods provide transparent and explainable results. Because the method requires strict control of the angle between the camera and surface normal, we integrate it with robotic handling and CAD‑based path planning to ensure scalable automation. The planner accounts for imaging constraints and variable processing time in areas with rapidly changing surface normals. This digital workflow also enables precise defect localization and supports cross‑process root‑cause analysis. Validated in a relevant environment, the system demonstrates effective digitization and inspection of complex composite surfaces, reducing operator workload and improving consistency. It supports greater digitalization in composite manufacturing, enhancing quality assurance, traceability, and productivity.