Generative Principal Component Thermography for Enhanced Defect Detection and Analysis

Machine learning methods play an important role in the nondestructive testing field for quality assessment of polymer composites. As a popular deep learning branch, a generative adversarial network is introduced to the thermography field as an image augmentation approach to improve its defect detection performance. Specifically, a generative principal component thermography (GPCT) method for defect detection in polymer composites is proposed. By employing the data augmentation strategy, more informative images are generated to enlarge the diversity of the original set of images. The defect detection results can be visualized using a number of interpretable features. Consequently, the defect detection performance of thermographic data analysis can be enhanced to some extent. The experimental results on a carbon fiber reinforced polymer specimen demonstrate the feasibility and advantages of the GPCT method.

[1]  Xavier Maldague,et al.  Nondestructive testing with thermography , 2013 .

[2]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[3]  Nik Rajic,et al.  Principal component thermography for flaw contrast enhancement and flaw depth characterisation in composite structures , 2002 .

[4]  Xavier Maldague,et al.  Solar loading thermography: Time-lapsed thermographic survey and advanced thermographic signal processing for the inspection of civil engineering and cultural heritage structures , 2017 .

[5]  Hayit Greenspan,et al.  GAN-based Synthetic Medical Image Augmentation for increased CNN Performance in Liver Lesion Classification , 2018, Neurocomputing.

[6]  A. Bendada,et al.  Comparative Study of Active Thermography Techniques for the Nondestructive Evaluation of Honeycomb Structures , 2009 .

[7]  Qi Xuan,et al.  Multiview Generative Adversarial Network and Its Application in Pearl Classification , 2019, IEEE Transactions on Industrial Electronics.

[8]  Chao Yang,et al.  Ensemble deep kernel learning with application to quality prediction in industrial polymerization processes , 2018 .

[9]  Hong Zhang,et al.  Smooth Nonnegative Matrix Factorization for Defect Detection Using Microwave Nondestructive Testing and Evaluation , 2014, IEEE Transactions on Instrumentation and Measurement.

[10]  Hongwei Liu,et al.  Convolutional Neural Network With Data Augmentation for SAR Target Recognition , 2016, IEEE Geoscience and Remote Sensing Letters.

[11]  X. Maldague,et al.  New absolute contrast for pulsed thermography , 2002 .

[12]  Krishnendu Chatterjee,et al.  Image Enhancement in Transient Lock-In Thermography Through Time Series Reconstruction and Spatial Slope Correction , 2012, IEEE Transactions on Instrumentation and Measurement.

[13]  Xavier Maldague,et al.  Low-rank sparse principal component thermography (sparse-PCT): Comparative assessment on detection of subsurface defects , 2019, Infrared Physics & Technology.

[14]  S. Gryś,et al.  Filtered thermal contrast based technique for testing of material by infrared thermography , 2011 .

[15]  Wai Lok Woo,et al.  Automatic Defect Identification of Eddy Current Pulsed Thermography Using Single Channel Blind Source Separation , 2014, IEEE Transactions on Instrumentation and Measurement.

[16]  Xavier Maldague,et al.  Improving the detection of thermal bridges in buildings via on-site infrared thermography: The potentialities of innovative mathematical tools , 2019, Energy and Buildings.

[17]  Xavier Maldague,et al.  Infrared thermography processing based on higher-order statistics , 2010 .

[18]  Yuan Yao,et al.  Stable principal component pursuit-based thermographic data analysis for defect detection in polymer composites , 2017 .

[19]  Amos J. Storkey,et al.  Data Augmentation Generative Adversarial Networks , 2017, ICLR 2018.

[20]  Biao Huang,et al.  Deep Learning-Based Feature Representation and Its Application for Soft Sensor Modeling With Variable-Wise Weighted SAE , 2018, IEEE Transactions on Industrial Informatics.

[21]  Wei-Min Liu,et al.  THE THERMOGRAPHIC SIGNAL RECONSTRUCTION METHOD: A POWERFUL TOOL FOR THE ENHANCEMENT OF TRANSIENT THERMOGRAPHIC IMAGES. , 2015, Biocybernetics and biomedical engineering.

[22]  Du-Ming Tsai,et al.  Automatic defect inspection for LCDs using singular value decomposition , 2005 .

[23]  Ling Shao,et al.  Zero-VAE-GAN: Generating Unseen Features for Generalized and Transductive Zero-Shot Learning , 2020, IEEE Transactions on Image Processing.

[24]  Yuan Yao,et al.  Online estimation and monitoring of local permeability in resin transfer molding , 2016 .

[25]  Soumith Chintala,et al.  Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks , 2015, ICLR.

[26]  Suneet Tuli,et al.  A comparison of the pulsed, lock-in and frequency modulated thermography nondestructive evaluation techniques , 2011 .

[27]  C. L. Philip Chen,et al.  GCB-Net: Graph Convolutional Broad Network and Its Application in Emotion Recognition , 2019, IEEE Transactions on Affective Computing.

[28]  L. Finesso,et al.  Matrix factorization methods: Application to thermal NDT/E , 2006 .

[29]  Yi Liu,et al.  Spatial-Neighborhood Manifold Learning for Nondestructive Testing of Defects in Polymer Composites , 2020, IEEE Transactions on Industrial Informatics.

[30]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[31]  Yuan Yao,et al.  Sparse Principal Component Thermography for Subsurface Defect Detection in Composite Products , 2018, IEEE Transactions on Industrial Informatics.