U-Net-Based Semantic Segmentation and Porosity Analysis of C/GFRP CT Images
DOI:
https://doi.org/10.54097/93ze7z97Keywords:
C/GFRP, CT, U-Net, void segmentation, layer porosity, damage evolutionAbstract
Accurate identification of internal voids and clarification of their evolution remain challenging in carbon/glass fiber hybrid composites (C/GFRP). In this study, interlaminar hybrid C/GFRP laminates were investigated using in situ tensile X-ray computed tomography (CT) to perform automatic void segmentation and quantitative analysis. CT images acquired at different loading stages were first preprocessed, and Gaussian, median, and non-local means (NLM) filters were compared, with NLM selected for denoising. Based on a finely annotated manual dataset, a U-Net semantic segmentation model was then established to automatically identify void regions and material-layer structures. On this basis, slice-wise layer-porosity variations along different section directions were analyzed, and the evolution of voids and cracks during loading was investigated in conjunction with three-dimensional reconstruction. The results show that the proposed segmentation workflow can stably identify voids and layer structures in C/GFRP CT images. Dice coefficients of 0.976 and 0.989 were obtained on the validation set for CFRP void segmentation and GFRP layer segmentation, respectively, indicating high accuracy and good generalization. The slice-wise layer porosity exhibits pronounced spatial non-uniformity in different section directions, and high-porosity regions persist during loading and correspond spatially to the final fracture location. Three-dimensional evolution results further reveal a damage process transitioning from dispersed voids to local concentration and then to crack coalescence and propagation. Damage in the GFRP layers propagates significantly faster than in the CFRP layers, indicating that the GFRP layers are the weak region governing progressive failure of the hybrid laminate. These findings provide methodological support for intelligent identification of internal defects, characterization of damage evolution, and correlation analysis of mechanical behavior in C/GFRP composites.
Downloads
References
[1] Rajak D K, Pagar D D, Menezes P L, et al. Fiber-Reinforced Polymer Composites: Manufacturing, Properties, and Applications [J]. Polymers, 2019, 11(10): 1667.
[2] Mehdikhani M, Gorbatikh L, Verpoest I, et al. Voids in fiber-reinforced polymer composites: A review on their formation, characteristics, and effects on mechanical performance [J]. Journal of Composite Materials, 2019, 53(12): 1579-1669.
[3] Dong C. Effects of Process-Induced Voids on the Properties of Fibre Reinforced Composites [J]. Journal of Materials Science & Technology, 2016, 32(7): 597-604.
[4] Liu X, Chen F. A review of void formation and its effects on the mechanical performance of carbon fiber reinforced plastic [J]. Engineering Transactions, 2016, 64(1): 33-51.
[5] Little J E, Yuan X, Jones M I. Characterisation of voids in fibre reinforced composite materials [J]. NDT & E International, 2012, 46: 122-127.
[6] Bodaghi M, Cristóvão C, Gomes R, et al. Experimental characterization of voids in high fibre volume fraction composites processed by high injection pressure RTM [J]. Composites Part A: Applied Science and Manufacturing, 2016, 82: 88-99.
[7] Patou J, Bonnaire R, De Luycker E, et al. Influence of consolidation process on voids and mechanical properties of powdered and commingled carbon/PPS laminates [J]. Composites Part A: Applied Science and Manufacturing, 2019, 117: 260-275.
[8] Tretiak I, Hallett S R. Predicting short beam shear strength reduction in carbon fibre composites containing voids [J]. Composite Structures, 2022, 290: 115472.
[9] Wang J, Zhou W, Wei Z Y, et al. Effects of various porosities on the damage evolution behavior of carbon fiber/epoxy composites using acoustic emission and micro-CT [J]. Journal of Composite Materials, 2022, 56(10): 1507-1524.
[10] Zhang J, Xie J, Zhao X, et al. Influence of void defects on impact properties of CFRP laminates based on multi-scale simulation method [J]. International Journal of Impact Engineering, 2023, 180: 104706.
[11] Jiang L, Xiong H, Zeng T, et al. In-situ micro-CT damage analysis of carbon and carbon/glass hybrid laminates under tensile loading by image reconstruction and DVC technology [J]. Composites Part A: Applied Science and Manufacturing, 2024, 176: 107844.
[12] Wang M, Zhang B, Zhan S, et al. Micro-Computed Tomography Non-Destructive Testing and Defect Quantitative Analysis of Carbon Fiber-Reinforced Polymer, Glass Fiber-Reinforced Polymer and Carbon/Glass Hybrid Laminates Using Deep Learning Recognition [J]. Applied Sciences, 2025, 15(22): 12192.
[13] Gao Y, Hu W, Xin S, et al. A review of applications of CT imaging on fiber reinforced composites [J]. Journal of Composite Materials, 2022, 56(1): 133-164.
[14] Naresh K, Khan K A, Umer R, et al. The use of X-ray computed tomography for design and process modeling of aerospace composites: A review [J]. Materials & Design, 2020, 190: 108553.
[15] Garcea S C, Wang Y, Withers P J. X-ray computed tomography of polymer composites [J]. Composites Science and Technology, 2018, 156: 305-319.
[16] Rashidi A, Olfatbakhsh T, Crawford B, et al. A Review of Current Challenges and Case Study toward Optimizing Micro-Computed X-Ray Tomography of Carbon Fabric Composites [J]. Materials, 2020, 13(16): 3606.
[17] Schilling P J, Karedla B P R, Tatiparthi A K, et al. X-ray computed microtomography of internal damage in fiber reinforced polymer matrix composites [J]. Composites Science and Technology, 2005, 65(14): 2071-2078.
[18] Gusenbauer C, Reiter M, Plank B, et al. Porosity Determination of Carbon and Glass Fibre Reinforced Polymers Using Phase-Contrast Imaging [J]. Journal of Nondestructive Evaluation, 2019, 38: 1.
[19] Zwanenburg E A, Norman D G, Qian C, et al. Effective X-ray micro computed tomography imaging of carbon fibre composites [J]. Composites Part B: Engineering, 2023, 258: 110707.
[20] Srivastava C, Agostino P, Stamopoulos A G, et al. Three-Dimensional Analysis of Porosity in As-Manufactured Glass Fiber/Vinyl Ester Filament Winded Composites Using X-Ray Micro-Computed Tomography [J]. Applied Composite Materials, 2024, 31: 171-200.
[21] Galvez-Hernandez P, Smith R A, Gaska K, et al. The effect of X-ray computed tomography scan parameters on porosity assessment of carbon fibre reinforced plastics laminates [J]. Journal of Composite Materials, 2023, 57(29): 4535-4548.
[22] Otsu N. A Threshold Selection Method from Gray-Level Histograms [J]. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1): 62-66.
[23] Canny J. A Computational Approach to Edge Detection [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1986, 8(6): 679-698.
[24] Tilton J C. Image Segmentation by Region Growing and Spectral Clustering with a Natural Convergence Criterion [C]//Proceedings of IGARSS 1998. Seattle, 1998.
[25] Levner I, Zhang H. Classification-driven watershed segmentation [J]. IEEE Transactions on Image Processing, 2007, 16(5): 1437-1445.
[26] Tretiak I, Smith R A. A parametric study of segmentation thresholds for X-ray CT porosity characterisation in composite materials [J]. Composites Part A: Applied Science and Manufacturing, 2019, 123: 10-19.
[27] Long J, Shelhamer E, Darrell T. Fully Convolutional Networks for Semantic Segmentation [C]//2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, 2015: 3431-3440.
[28] Ronneberger O, Fischer P, Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation [C]//Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015. Cham: Springer, 2015: 234-241.
[29] Chen L C, Barron J T, Papandreou G, et al. Semantic Image Segmentation with Task-Specific Edge Detection Using CNNs and a Discriminatively Trained Domain Transform [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, 2016: 4545-4554.
[30] Chen L C, Zhu Y, Papandreou G, et al. Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation [C]//Computer Vision – ECCV 2018. Cham: Springer, 2018: 833-851.
[31] Zhou Z, Siddiquee M M R, Tajbakhsh N, et al. UNet++: A Nested U-Net Architecture for Medical Image Segmentation [C]//Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support. Cham: Springer, 2018: 3-11.
[32] Sinchuk Y, Kibleur P, Aelterman J, et al. Variational and Deep Learning Segmentation of Very-Low-Contrast X-ray Computed Tomography Images of Carbon/Epoxy Woven Composites [J]. Materials, 2020, 13(4): 936.
[33] Galvez Hernandez P, Gaska K, Kratz J. Phase segmentation of uncured prepreg X-Ray CT micrographs [J]. Composites Part A: Applied Science and Manufacturing, 2021, 149: 106527.
[34] Zhong Q, Zhang J, Xu Y, et al. Filamentous target segmentation of weft micro-CT image based on U-Net [J]. Micron, 2021, 146: 103072.
[35] Machado J M, Tavares J M R S, Camanho P P, et al. Automatic void content assessment of composite laminates using a machine-learning approach [J]. Composite Structures, 2022, 288: 115383.
[36] Badran A, Parkinson D, Ushizima D, et al. Validation of Deep Learning Segmentation of CT Images of Fiber-Reinforced Composites [J]. Journal of Composites Science, 2022, 6(2): 60.
[37] Helwing R, Hülsbusch D, Walther F. Deep learning method for analysis and segmentation of fatigue damage in X-ray computed tomography data for fiber-reinforced polymers [J]. Composites Science and Technology, 2022, 230: 109781.
[38] Upadhyay S, George Smith A, Vandepitte D, et al. Deep-learning versus greyscale segmentation of voids in X-ray computed tomography images of filament-wound composites [J]. Composites Part A: Applied Science and Manufacturing, 2024, 177: 107937.
[39] Petkov V I, Pakkam Gabriel V R, Fernberg P. Semantic segmentation of progressive micro-cracking in polymer composites using Attention U-Net architecture [J]. Tomography of Materials and Structures, 2024, 5: 100028.
[40] Zheng K, Cao X, Jiang Z, et al. Improved XCT image automatic segmentation for quantitative characterization of the meso-morphological features in the damaged braided composite fabric [J]. Composites Science and Technology, 2024, 247: 110395.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Academic Journal of Applied Sciences
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
