Improving visibility of rear surface cracks during inductive thermography of metal plates using Autoencoder

Abstract Inductive thermography is one kind of infrared thermography (IRT) technique, which is effective in detection of front surface cracks in metal plates. However, rear surface cracks are usually missed due to their weak indications during inductive thermography. Here we propose a novel approach (AET: AE Thermography) to improve the visibility of rear surface cracks during inductive thermography by employing the Autoencoder (AE) algorithm, which is an important block to construct deep learning architectures. We construct an integrated framework for processing the raw inspection data of inductive thermography using the AE algorithm. Through this framework, underlying features of rear surface cracks are efficiently extracted and new clearer images are constructed. Experiments of inductive thermography were conducted on steel specimens to verify the efficacy of the proposed approach. We visually compare the raw thermograms, the empirical orthogonal functions (EOFs) of the prominent component thermography (PCT) technique and the results of AET. We further quantitatively evaluated AET by calculating crack contrast and signal-to-noise ratio (SNR). The results demonstrate that the proposed AET approach can remarkably improve the visibility of rear surface cracks and then improve the capability of inductive thermography in detecting rear surface cracks in metal plates.

[1]  Yunze He,et al.  Eddy current pulsed phase thermography for subsurface defect quantitatively evaluation , 2013 .

[2]  Ermanno G. Grinzato,et al.  Statistical analysis of IR thermographic sequences by PCA , 2004 .

[3]  Vladimir P. Vavilov,et al.  Crack detection in aluminum parts by using ultrasound-excited infrared thermography , 2013 .

[4]  Geoffrey E. Hinton,et al.  On the importance of initialization and momentum in deep learning , 2013, ICML.

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

[6]  Pascal Vincent,et al.  Stacked Denoising Autoencoders: Learning Useful Representations in a Deep Network with a Local Denoising Criterion , 2010, J. Mach. Learn. Res..

[7]  Yoshua Bengio,et al.  What regularized auto-encoders learn from the data-generating distribution , 2012, J. Mach. Learn. Res..

[8]  Dan Zhang,et al.  Stacked Sparse Autoencoder in PolSAR Data Classification Using Local Spatial Information , 2016, IEEE Geoscience and Remote Sensing Letters.

[9]  Tetsuya Ogata,et al.  Audio-visual speech recognition using deep learning , 2014, Applied Intelligence.

[10]  Simon J. Doran,et al.  Stacked Autoencoders for Unsupervised Feature Learning and Multiple Organ Detection in a Pilot Study Using 4D Patient Data , 2013, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[11]  Lihi Zelnik-Manor,et al.  Context-Aware Saliency Detection , 2012, IEEE Trans. Pattern Anal. Mach. Intell..

[12]  Xavier Maldague,et al.  Theory and Practice of Infrared Technology for Nondestructive Testing , 2001 .

[13]  Thierry Palin-Luc,et al.  Fatigue crack initiation and growth in a 35CrMo4 steel investigated by infrared thermography , 2005 .

[14]  Björn W. Schuller,et al.  Autoencoder-based Unsupervised Domain Adaptation for Speech Emotion Recognition , 2014, IEEE Signal Processing Letters.

[15]  Geoffrey E. Hinton,et al.  Reducing the Dimensionality of Data with Neural Networks , 2006, Science.

[16]  Geoffrey E. Hinton,et al.  Modeling the manifolds of images of handwritten digits , 1997, IEEE Trans. Neural Networks.

[17]  Darryl P Almond,et al.  Crack imaging by scanning laser-line thermography and laser-spot thermography , 2011 .

[18]  Xavier Maldague,et al.  Pulsed Thermography Signal Processing Techniques Based on the 1D Solution of the Heat Equation Applied to the Inspection of Laminated Composites , 2014 .

[19]  Paul O'Leary,et al.  Thermographic crack detection and failure classification , 2010, J. Electronic Imaging.

[20]  Takahide Sakagami,et al.  Development of self-reference lock-in thermography and its application to remote nondestructive inspection of fatigue cracks in steel bridges , 2010 .

[21]  Xavier Maldague,et al.  Optimization of pulsed thermography inspection by partial least-squares regression , 2014 .

[22]  Xavier Maldague,et al.  Advances in pulsed phase thermography , 2002 .

[23]  G. Zenzinger,et al.  Crack Detection Using EddyTherm , 2005 .

[24]  Julio Molleda,et al.  Infrared Thermography for Temperature Measurement and Non-Destructive Testing , 2014, Sensors.

[25]  W. Arnold,et al.  MECHANISMS AND MODELS FOR CRACK DETECTION WITH INDUCTION THERMOGRAPHY , 2008 .

[26]  P. Cawley,et al.  Eddy-current induced thermography—probability of detection study of small fatigue cracks in steel, titanium and nickel-based superalloy , 2012 .

[27]  Nasser M. Nasrabadi,et al.  Pattern Recognition and Machine Learning , 2006, Technometrics.

[28]  Bruce A. Rubadeux,et al.  Reconstruction and enhancement of active thermographic image sequences , 2003 .

[29]  Norbert Meyendorf,et al.  Simulation of the surface crack detection using inductive heated thermography , 2012 .

[30]  U. Netzelmann,et al.  Induction thermography: principle, applications and first steps towards standardisation , 2016 .