Automated detection of edge trimming-induced delamination in GFRPs using SIFT-based image registration and differencing

Abstract

Abstract Fibre-reinforced polymer (FRP) composites are increasingly used in aerospace and automotive applications due to their high strength and corrosion resistance, but their abrasive, heterogeneous structure makes post-processing difficult. A major machining defect is edge trimming-induced delamination, which affects both appearance and structural integrity. Existing measurement methods are often manual, subjective, and not fully reproducible. This study proposes an automated, reproducible method for measuring delamination using SIFT-based image registration and image differencing. The approach compares “before” and “after” images of FRP specimens, aligns them with the SIFT algorithm, and detects damage through pixel intensity changes. Validation was performed on glass fibre reinforced polymer (GFRP) composite specimens machined with a Ø10 mm FRAISA compression end mill and imaged using a telecentric Mitutoyo QIA505 digital microscope. Results were compared with semi-manual measurements. The proposed method showed excellent performance, with an accuracy of 0.998 ± 0.001, precision of 0.763 ± 0.109, and recall of 0.858 ± 0.094. Its speed and objectivity make it suitable for industrial quality control and large-scale composite engineering studies.