Infrared-Induced Laser Shearography: Enhanced Multimodal Features Recognition for Interfacial Defects in SIR/GFRP Composite Structures

The detection capability of laser shearography has not been fully developed due to neglecting the overall dynamic deformation process. In this article, we propose an infrared-induced laser shearography (IILS) method for composite structure detection that captures dynamic deformation data and enhances detection performance through feature extraction. Based on the temperature-deformation connection theory, we investigate the relationship between the dynamic deformation process and defects through simulations and experiments, demonstrating that the dynamic process contains a wealth of defect information. Using feature extraction algorithms, we perform denoising and enhance imaging of the shearogram sequence. Applying tensor robust principal component analysis (TRPCA), we extract a low tubal rank tensor that captures the intrinsic deformations and performs additional feature extraction. We validate the benefits of the proposed TRPCA-based feature extraction method in image enhancement through visual observation and evaluation using parameters such as two-dimensional entropy (TDE), contrast, and signal-to-noise ratio (SNR). In conclusion, the dynamic deformation-based IILS method significantly enhances the detection capability, reliability, and stability of interferometric modality in the field of industrial non-destructive testing.