Unsupervised Learning for Industrial Defect Detection: A Case Study on Shearographic Data

TL;DR

This paper investigates unsupervised deep learning methods, including autoencoders and a student-teacher model, for automated defect detection in shearographic images, aiming to reduce reliance on labeled data in industrial inspections.

Contribution

It introduces and evaluates three unsupervised architectures for defect detection in shearography, demonstrating the superior performance of a student-teacher model in classification and localization tasks.

Findings

Student-teacher model achieves better classification robustness.

Model enables precise defect localization.

Unsupervised methods outperform autoencoders in feature separation.

Abstract

Shearography is a non-destructive testing method for detecting subsurface defects, offering high sensitivity and full-field inspection capabilities. However, its industrial adoption remains limited due to the need for expert interpretation. To reduce reliance on labeled data and manual evaluation, this study explores unsupervised learning methods for automated anomaly detection in shearographic images. Three architectures are evaluated: a fully connected autoencoder, a convolutional autoencoder, and a student-teacher feature matching model. All models are trained solely on defect-free data. A controlled dataset was developed using a custom specimen with reproducible defect patterns, enabling systematic acquisition of shearographic measurements under both ideal and realistic deformation conditions. Two training subsets were defined: one containing only undistorted, defect-free samples, and one additionally including globally deformed, yet defect-free, data. The latter simulates practical inspection conditions by incorporating deformation-induced fringe patterns that may obscure localized anomalies. The models are evaluated in terms of binary classification and, for the student-teacher model, spatial defect localization. Results show that the student-teacher approach achieves superior classification robustness and enables precise localization. Compared to the autoencoder-based models, it demonstrates improved separability of feature representations, as visualized through t-SNE embeddings. Additionally, a YOLOv8 model trained on labeled defect data serves as a reference to benchmark localization quality. This study underscores the potential of unsupervised deep learning for scalable, label-efficient shearographic inspection in industrial environments.