基于深度学习的复合材料层合板损伤图像分类的研究
Research on Damage Image Classification of Composite Laminates Based on Deep Learning
DOI: 10.12677/CSA.2024.142031, PDF,    国家自然科学基金支持
作者: 王正水, 吴慧婕, 孙汤慧：南昌航空大学数学与信息科学学院，江西 南昌；赵 刚*：南昌航空大学数学与信息科学学院，江西 南昌;南昌航空大学无损检测技术教育部重点实验室，江西 南昌
关键词: 复合材料；卷积神经网络；损伤检测；深度学习；Composite Material； Convolutional Neural Network； Damage Detection； Deep Learning
摘要: 针对复合材料结构检测损伤检测问题，本文提出了一种基于深度学习进行复合材料结构损伤检测的方法。本方法首先通过网络和文献收集复合材料结构图像资料，建立数据集，数据集包含损伤和未损伤的复合材料层合板图片；然后采用三个卷积神经网络模型AlexNet、VGG和ResNet对损伤情况进行自动分类；最后对三种预先训练过的网络架构的性能进行评估。实验结果表明，在相同的实验条件下，AlexNet技术使用相对较小的图像数据集，在合理的计算时间内能够成功地检测出损伤，且测试精度最高，复杂性较低。
Abstract: Composite structure detection technology has been exploring the efficient and fast damage detec-tion technology. In this paper, an image-based NDT technique is proposed to detect composite ma-terial damage by deep learning. A dataset containing damaged and non-damaged composite laminate images was established through the network and literature data. Then three convolutional neural network models AlexNet, VGG and ResNet were used to automatically classify the damage conditions. Finally, the performance of three pretrained network architectures is evaluated. The results show that AlexNet technology can successfully detect damage within a reasonable calculation time using a relatively small image dataset, with the highest test accuracy and low complexity.
文章引用：王正水, 赵刚, 吴慧婕, 孙汤慧. 基于深度学习的复合材料层合板损伤图像分类的研究[J]. 计算机科学与应用, 2024, 14(2): 308-316. https://doi.org/10.12677/CSA.2024.142031

参考文献

[1] Abramovich, H. (2017) Introduction to Composite Materials. Woodhead Publishing, England, 1-47. [Google Scholar] [CrossRef] 
[2] Fernandes, H., Zhang, H., Quirin, S., et al. (2022) In-frared Thermographic Inspection of 3D Hybrid Aluminium-CFRP Composite Using Different Spectral Bands and New Unsupervised Probabilistic Low-Rank Component Factorization Model. NDT & E International, 125, Article ID: 102561. [Google Scholar] [CrossRef] 
[3] Fotouhi, M. and Ahmadi Najafabadi, M. (2014) Inves-tigation of the Mixed-Mode Delamination in Polymer-Matrix Composites Using Acoustic Emission Technique. Journal of Reinforced Plastics and Composites, 33, 1767-1782. [Google Scholar] [CrossRef] 
[4] Gholizadeh, S. (2016) A Review of Non-Destructive Testing Methods of Composite Materials. Procedia Structural Integrity, 1, 50-57. [Google Scholar] [CrossRef] 
[5] Jenssen, R. and Roverso, D. (2018) Automatic Autonomous Vi-sion-Based Power Line Inspection: A Review of Current Status and the Potential Role of Deep Learning. International Journal of Electrical Power & Energy Systems, 99, 107-120. [Google Scholar] [CrossRef] 
[6] Ma, A.M., Yu, B.J., Fan, C.W., et al. (2022) Damage Detection of Carbon Fiber Reinforced Polymer Composite Materials Based on One-Dimensional Multi-Scale Residual Convolution Neural Network. Review of Scientific Instruments, 93, Ar-ticle ID: 034701. [Google Scholar] [CrossRef] [PubMed]
[7] Ravandi, M., Teo, W.S., Tran, L.Q.N., et al. (2017) Low Velocity Impact Performance of Stitched Flax/Epoxy Composite Laminates. Composites Part B: Engineering, 117, 89-100. [Google Scholar] [CrossRef] 
[8] Fu, H., Feng, X., Liu, J., et al. (2020) An Investi-gation on Anti-Impact and Penetration Performance of Basalt Fiber Composites with Different Weave and Lay-up Modes. Defence Technology, 16, 787-801. [Google Scholar] [CrossRef] 
[9] Shohag, M.A.S., Hammel, E.C., Olawale, D.O., et al. (2017) Damage Mitigation Techniques in Wind Turbine Blades: A Review. Wind Engineering, 41, 185-210. [Google Scholar] [CrossRef] 
[10] Fotouhi, S., Pashmforoush, F., Bodaghi, M., et al. (2021) Au-tonomous Damage Recognition in Visual Inspection of Laminated Composite Structures Using Deep Learning. Compo-site Structures, 268, Article ID: 113960. [Google Scholar] [CrossRef] 
[11] Tang, E., Wang, J., Han, Y., et al. (2019) Microscopic Damage Modes and Physical Mechanisms of CFRP Laminates Impacted by Ice Projectile at High Velocity. Journal of Materials Research and Technology, 8, 5671-5686. [Google Scholar] [CrossRef] 
[12] Wang, Z., Yang, J., Jiang, H., et al. (2020) CNN training with Twenty Samples for Crack Detection via Data Augmentation. Sensors, 20, Article 4849. [Google Scholar] [CrossRef] [PubMed]
[13] Zhao, X.Y., Dong, C.Y., Zhou, P., et al. (2019) Detecting Surface Defects of Wind Tubine Blades Using an Alexnet Deep Learning Algorithm. Ieicetransactions on Fundamentals of Electronics, Communications and Computer Sciences, 102, 1817-1824. [Google Scholar] [CrossRef] 
[14] Theckedath, D. and Sedamkar, R.R. (2020) Detecting Affect States Using VGG16, ResNet50 and SE-ResNet50 Networks. Sncomputer Science, 1, Article No. 79. [Google Scholar] [CrossRef] 
[15] Mahajan, A. and Chaudhary, S. (2019) Categorical Image Classi-fication Based on Representational Deep Network (RESNET). 2019 3rd International Conference on Electronics, Com-munication and Aerospace Technology (ICECA), Coimbatore, 12-14 June 2019, 327-330. [Google Scholar] [CrossRef]