Detectability enhancement of small defects in materials with high structural noise: A deep learning approach for baseline signal reconstruction

Abstract

This paper investigates the suppression of background noise in ultrasonic array imaging by applying the total focusing method and baseline subtraction, focusing on coarse-grained materials that exhibit significant levels of structural noise. Addressing the challenge of identifying small defects due to low signal-to-noise ratios (SNRs) in the measured array data, we have proposed an efficient methodology that can be applied to enhance the detectability of a defect within a specified region of interest (ROI). The proposed methodology requires the original full matrix capture data solely, and it generates the reconstructed baseline (i.e., estimated grain noise) data using a multi-step long short-term memory model. This model predicts time traces corresponding to the ROI based on historical signals of the same data set. The root mean square value and peak noise amplitude of the reconstructed grain image are used to evaluate the noise prediction performance of the proposed approach. The simulation and experimental study results demonstrate that our proposed approach for reconstructing the grain-scattered data can notably enhance the defect SNR when combined with a straightforward baseline subtraction method. Moreover, the effect of the probe position (with respect to the defect) on the noise suppression capability of the prediction model is shown to be small.

Graphical abstract

Keywords

ultrasonic arrays / structural noise / baseline subtraction / total focusing method

References

[1] Zou X , Gao W , Liu G . Low-velocity impact damage detection in CFRP laminates based on ultrasonic phased-array NDT technique. Russian Journal of Nondestructive Testing, 2023, 59(8): 876–885

[2] Zhuang B , Gencturk B , Sinkov A , Good M , Meyer R , Oberai A . Non-invasive ultrasonic sensing of internal conditions on a partial full-scale spent nuclear fuel canister mock-up. NDT & E International, 2024, 148: 103242

[3] Zhang J , Song Y , Li X , Zhong C H . Comparison of experimental measurements of material grain size using ultrasound. Journal of Nondestructive Evaluation, 2020, 39(2): 30

[4] Yu Z , Chen J , Wu S , Xie Y , Wu H , Wang H , Peng H X . Adaptive ultrasonic full-matrix imaging of internal defects in CFRP laminates with arbitrary stacking sequences. Composites Part B: Engineering, 2024, 275: 111309

[5] Nicolson E , Mohseni E , Lines D , Tant K M M , Pierce G , MacLeod C N . Towards an in-process ultrasonic phased array inspection method for narrow-gap welds. NDT & E International, 2024, 144: 103074

[6] Schmerr L W. Fundamentals of Ultrasonic Phased Arrays. Cham Heidelberg: Springer, 2015

[7] Drinkwater B W , Wilcox P D . Ultrasonic arrays for non-destructive evaluation: A review. NDT & E International, 2006, 39(7): 525–541

[8] Montaldo G , Tanter M , Bercoff J , Benech N , Fink M . Coherent plane-wave compounding for very high frame rate ultrasonography and transient elastography. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2009, 56(3): 489–506

[9] Holmes C , Drinkwater B W , Wilcox P D . Post-processing of the full matrix of ultrasonic transmit–receive array data for non-destructive evaluation. NDT & E International, 2005, 38(8): 701–711

[10] Zhang J , Drinkwater B W , Wilcox P D . Defect characterization using an ultrasonic array to measure the scattering coefficient matrix. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2008, 55(10): 2254–2265

[11] Hunter A J , Drinkwater B W , Wilcox P D . The wavenumber algorithm for full-matrix imaging using an ultrasonic array. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2008, 55(11): 2450–2462

[12] Velichko A , Wilcox P D . Reversible back-propagation imaging algorithm for post processing of ultrasonic array data. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2009, 56(11): 2492–2503

[13] Velichko A , Wilcox P D . An analytical comparison of ultrasonic array imaging algorithms. Journal of the Acoustical Society of America, 2010, 127(4): 2377–2384

[14] Felice M V , Fan Z . Sizing of flaws using ultrasonic bulk wave testing: A review. Ultrasonics, 2018, 88: 26–42

[15] Li C , Wang R , Liu Z , Ji H , Qin S , Chen R , Zhong Y , Wu H . Study on phased array ultrasonic testing techniques for austenitic stainless steel butt welds. Journal of Fusion Energy, 2025, 44(1): 13

[16] Luo Z , Kang J , Cao H , Lin L . Enhanced ultrasonic total focusing imaging of CFRP corner with ray theory-based homogenization technique. Chinese Journal of Aeronautics, 2023, 36(1): 434–443

[17] Selim H , Delgado-Prieto M , Trull J , Picó R , Romeral L , Cojocaru C . Defect reconstruction by non-destructive testing with laser induced ultrasonic detection. Ultrasonics, 2020, 101: 106000

[18] Haupert S , Ohara Y , Carcreff E , Renaud G . Fundamental wave amplitude difference imaging for detection and characterization of embedded cracks. Ultrasonics, 2019, 96: 132–139

[19] Saini A , Lane C J L , Tu J , Xue H , Fan Z . 3D ultrasonic imaging of surface-breaking cracks using a linear array. Ultrasonics, 2022, 125: 106790

[20] Nicolas P , Paul K , Antoine F , Andreas S , Pierre-Emile L . Ultrasound array probe: Signal processing in case of structural noise. Journal of Nondestructive Evaluation, Diagnostics and Prognostics of Engineering Systems, 2020, 3(4): 041003

[21] Newhouse V L , Bilgutay N M , Saniie J , Furgason E S . Flaw-to-grain echo enhancement by split-spectrum processing. Ultrasonics, 1982, 20(2): 59–68

[22] Tian Q , Bilgutay N M . Statistical analysis of split spectrum processing for multiple target detection. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 1998, 45(1): 251–256

[23] Pedram S K , Fateri S , Gan L , Haig A , Thornicroft K . Split-spectrum processing technique for SNR enhancement of ultrasonic guided wave. Ultrasonics, 2018, 83: 48–59

[24] Bai L , Velichko A , Drinkwater B W . Grain scattering noise modeling and its use in the detection and characterization of defects using ultrasonic arrays. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2019, 66(11): 1798–1813

[25] Matz V , Smid R , Starman S , Kreidl M . Signal-to-noise ratio enhancement based on wavelet filtering in ultrasonic testing. Ultrasonics, 2009, 49(8): 752–759

[26] Pardo E , San Emeterio J L , Rodriguez M A , Ramos A . Noise reduction in ultrasonic NDT using undecimated wavelet transforms. Ultrasonics, 2006, 44: e1063–e1067

[27] Bettayeb F , Haciane S , Aoudia S . Improving the time resolution and signal noise ratio of ultrasonic testing of welds by the wavelet packet. NDT & E International, 2005, 38(6): 478–484

[28] Xu W , Li X , Zhang J , Xue Z , Cao J . Ultrasonic signal enhancement for coarse grain materials by machine learning analysis. Ultrasonics, 2021, 117: 106550

[29] Munir M , Kim H J , Park J , Song S J , Kang S S . Convolutional neural network for ultrasonic weldment flaw classification in noisy conditions. Ultrasonics, 2019, 94: 74–81

[30] Roca Barceló F , Jaén Del Hierro P , Ribes Llario F , Real Herráiz J . Development of an ultrasonic weld inspection system based on image processing and neural networks. Nondestructive Testing and Evaluation, 2018, 33(2): 229–236

[31] Cantero-Chinchilla S , Wilcox P D , Croxford A J . Deep learning in automated ultrasonic NDE – developments, axioms and opportunities. NDT & E International, 2022, 131: 102703

[32] Liu Y , Yang Z , Zou X , Ma S , Liu D , Avdeev M , Shi S . Data quantity governance for machine learning in materials science. National Science Review, 2023, 10(7): nwad125

[33] Liu Y , Ma S , Yang Z , Zou X , Shi S . A Data quality and quantity governance for machine learning in materials science. Journal of the Chinese Ceramic Society, 2023, 51: 427–437

[34] Liu Y , Guo B , Zou X , Li Y , Shi S . Machine learning assisted materials design and discovery for rechargeable batteries. Energy Storage Materials, 2020, 31: 434–450

[35] Jiang X , Wang W , Tian S , Wang H , Lookman T , Su Y . Applications of natural language processing and large language models in materials discovery. npj Computational Materials, 2025, 11(1): 79

[36] Jain A , Ong S P , Hautier G , Chen W , Richards W D , Dacek S , Cholia S , Gunter D , Skinner D , Ceder G , Persson K A . Commentary: The materials project. A materials genome approach to accelerating materials innovation. APL Materials, 2013, 1(1): 011002

[37] Liu Y , Ding L , Yang Z , Ge X , Liu D , Liu W , Yu T , Avdeev M , Shi S . Domain knowledge discovery from abstracts of scientific literature on nickel-based single crystal superalloys. Science China Technological Sciences, 2023, 66(6): 1815–1830

[38] Liu Y , Wu L , Yang Z , Zou X , Zou Z , Lin Y , Avdeev M , Shi S . Descriptors divide-and-conquer enables multifaceted and interpretable materials structure–activity relationship analysis. Advanced Functional Materials, 2025, 35(26): 2421621

[39] Mariani S , Heinlein S , Cawley P . Location specific temperature compensation of guided wave signals in structural health monitoring. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2020, 67(1): 146–157

[40] Dawson A J, Michaels J E, Michaels T E. Isolation of ultrasonic scattering by wavefield baseline subtraction. Mechanical Systems and Signal Processing, 2016, 70–71: 891–903

[41] Xue Y , Croxford A J , Wilcox P D . Improvement in repeatability of ultrasonic array imaging in materials with high structural noise. NDT & E International, 2023, 133: 102732

[42] Ye T , Zhang J , Du Y , Xu J , Bai L . A receiver-optimized total focusing method for detectability enhancement of small defects in coarse grained materials. NDT & E International, 2023, 140: 102943

[43] Hochreiter S , Schmidhuber J . Long short-term memory. Neural Computation, 1997, 9(8): 1735–1780

[44] Wilcox P D . Ultrasonic arrays in NDE: Beyond the B-scan. AIP Conference Proceedings, 2013, 1511: 33–50

[45] Aurenhammer F . Voronoi diagrams—a survey of a fundamental geometric data structure. ACM Computing Surveys, 1991, 23(3): 345–405

[46] Huthwaite P . Accelerated finite element elastodynamic simulations using the GPU. Journal of Computational Physics, 2014, 257: 687–707

[47] Van Pamel A , Brett C R , Huthwaite P , Lowe M J S . Finite element modelling of elastic wave scattering within a polycrystalline material in two and three dimensions. Journal of the Acoustical Society of America, 2015, 138(4): 2326–2336

[48] Medsker L, Jain L. Recurrent Neural Networks: Design and Applications. Boca Raton: CRC Press, 1999

[49] Cambria E , White B . Jumping NLP curves: A review of natural language processing research. IEEE Computational Intelligence Magazine, 2014, 9(2): 48–57

[50] Sezer O B , Gudelek M U , Ozbayoglu A M . Financial time series forecasting with deep learning: A systematic literature review: 2005–2019. Applied Soft Computing, 2020, 90: 106181

[51] Hua Y , Zhao Z , Li R , Chen X , Liu Z , Zhang H . Deep learning with long short-term memory for time series prediction. IEEE Communications Magazine, 2019, 57(6): 114–119

[52] Box G E P, Jenkins G M, Reinsel G C, Ljung G M. Time Series Analysis: Forecasting and Control. 5th ed. Hoboken: John Wiley & Sons, 2015

[53] Toscano J D , Oommen V , Varghese A J , Zou Z , Ahmadi Daryakenari N , Wu C , Karniadakis G E . From PINNs to PIKANs: recent advances in physics-informed machine learning. Machine Learning for Computational Science and Engineering, 2025, 1: 15

[54] Pyle R J , Bevan R L T , Hughes R R , Ali A A S , Wilcox P D . Domain adapted deep-learning for improved ultrasonic crack characterization using limited experimental data. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2022, 69(4): 1485–1496

[55] Lim B , Arık S Ö , Loeff N , Pfister T . Temporal fusion transformers for interpretable multi-horizon time series forecasting. International Journal of Forecasting, 2021, 37(4): 1748–1764

[56] Liu Y , Yang Z , Yu Z , Liu Z , Liu D , Lin H , Li M , Ma S , Avdeev M , Shi S . Generative artificial intelligence and its applications in materials science: Current situation and future perspectives. Journal of Materiomics, 2023, 9(4): 798–816

RIGHTS & PERMISSIONS

Higher Education Press