Abstract
Detection of fiber composite material boundaries and defects is critical to the automation of the manufacturing process in the aviation industry. This paper describes a process to evaluate four well-performing deep convolutional neural network models (Mask R-CNN, U-Net, DeepLab V3+, and IC-Net) for use in such a process. A custom-captured dataset of images showing fiber cut-pieces with geometrical defects was annotated and augmented for training deep convolutional neural network models; results show acceptable detection accuracy for gripper and fabric based on the Intersection over Union (IoU) scores of up to 0.92 and 0.86, respectively. However, wrinkle detection initially achieves a significantly lower IoU score of 0.40 in the best case. This discrepancy is mainly due to geometrical ambiguities, as wrinkles do not have a clearly defined boundary and are hard to distinguish even for human eye. The model is then evaluated as a binary predictor based on per-component detection success; the model achieves a recall rate (i.e., the ratio of the wrinkles detected to all existing wrinkles) of 0.71 and a precision score (i.e., the ratio of those detected being actually wrinkles) of 0.76. From a practical point of view, this model can outperform a human operator based on the results presented. Two complementary approaches are also introduced for the detection of wrinkles at the early stages of formation as well as the completely formed wrinkles. The developed method can be readily used in a variety of composite manufacturing processes or adapted to other similar tasks.
Similar content being viewed by others
Notes
Intersection over Union.
References
Abdul Ghafour, T., Colmars, J., & Boisse, P. (2019). The importance of taking into account behavior irreversibilities when simulating the forming of textile composite reinforcements. Composites Part A: Applied Science and Manufacturing. https://doi.org/10.1016/j.compositesa.2019.105641
Alshahrani, H., & Hojjati, M. (2017). A new test method for the characterization of the bending behavior of textile prepregs. Composites Part A: Applied Science and Manufacturing, 97(97), 128–140. https://doi.org/10.1016/j.compositesa.2017.02.027
Anantharaman, R., Velazquez, M., & Lee, Y. (2018). Utilizing mask R-CNN for detection and segmentation of oral diseases. In 2018 IEEE international conference on bioinformatics and biomedicine (BIBM) (pp. 2197–2204).
Aridhi, A., Arfaoui, M., Mabrouki, T., Naouar, N., Denis, Y., Zarroug, M., & Boisse, P. (2019). Textile composite structural analysis taking into account the forming process. Composites Part B: Engineering, 166, 773–784. https://doi.org/10.1016/j.compositesb.2019.02.047
Bian, X., Lim, S. N., & Zhou, N. (2016). Multiscale fully convolutional network with application to industrial inspection. In 2016 IEEE winter conference on applications of computer vision, WACV 2016. https://doi.org/10.1109/WACV.2016.7477595
Boisse, P., Colmars, J., Hamila, N., Naouar, N., & Steer, Q. (2018). Bending and wrinkling of composite fiber preforms and prepregs. A review and new developments in the draping simulations. Composites Part B: Engineering. https://doi.org/10.1016/j.compositesb.2017.12.061
Cao, J., Akkerman, R., Boisse, P., Chen, J., Cheng, H. S., De Graaf, E. F., Gorczyca, J. L., Harrison, P., Hivet, G., Launay, J., & Lee, W. (2008). Characterization of mechanical behavior of woven fabrics: Experimental methods and benchmark results. Composites Part A: Applied Science and Manufacturing, 39(6), 1037–1053. https://doi.org/10.1016/j.compositesa.2008.02.016
Chen, L.-C., Papandreou, G., Kokkinos, I., Murphy, K., & Yuille, A. L. (2017a). Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs. IEEE Transactions on Pattern Analysis and Machine Intelligence, 40(4), 834–848
Chen, L.-C., Papandreou, G., Schroff, F., & Adam, H. (2017). Rethinking atrous convolution for semantic image segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition.
Chen, L.-C., Zhu, Y., Papandreou, G., Schroff, F., & Adam, H. (2018). Encoder-decoder with atrous separable convolution for semantic image segmentation. In Proceedings of the European conference on computer vision (ECCV) (pp. 801–818).
Dangora, L. M., Mitchell, C., White, K. D., Sherwood, J. A., & Parker, J. C. (2018). Characterization of temperature-dependent tensile and flexural rigidities of a cross-ply thermoplastic lamina with implementation into a forming model. International Journal of Material Forming, 11(1), 43–52. https://doi.org/10.1007/s12289-016-1327-2
Dempster, A. P. (2008). Upper and lower probabilities induced by a multivalued mapping. In Classic works of the dempster-shafer theory of belief functions (pp. 57–72). Springer.
Denis, Y., Morestin, F., & Hamila, N. (2020). A hysteretic model for fiber-reinforced composites at finite strains: Fractional derivatives, computational aspects and analysis. Computational Materials Science, 181, 109716. https://doi.org/10.1016/j.commatsci.2020.109716
Djavadifar, A., Graham-Knight, J. B., Gupta, K., Körber, M., Lasserre, P., & Najjaran, H. (2019). Robot-assisted composite manufacturing based on machine learning applied to multi-view computer vision. In International conference on smart multimedia.
Doitrand, A., Fagiano, C., Irisarri, F. X., & Hirsekorn, M. (2015). Comparison between voxel and consistent meso-scale models of woven composites. Composites Part A: Applied Science and Manufacturing, 73(73), 143–154. https://doi.org/10.1016/j.compositesa.2015.02.022
Dong, X., Taylor, C. J., & Cootes, T. F. (2019). Small defect detection using convolutional neural network features and random forests. In Lecture notes in computer science (including subseries lecture notes in artificial intelligence and lecture notes in bioinformatics), 11132 LNCS (pp. 398–412). https://doi.org/10.1007/978-3-030-11018-5_35
Enshaei, N., Ahmad, S., & Naderkhani, F. (2020). Automated detection of textured-surface defects using UNet-based semantic segmentation network. In Proceedings of the annual conference of the prognostics and health management society, PHM, 2020-June. https://doi.org/10.1109/ICPHM49022.2020.9187023
Everingham, M., Van Gool, L., Williams, C. K. I., Winn, J., & Zisserman, A. (2010). The pascal visual object classes (VOC) challenge. International Journal of Computer Vision, 88(2), 303–338. https://doi.org/10.1007/s11263-009-0275-4
Faal, R. T., Sourki, R., Crawford, B., Vaziri, R., & Milani, A. S. (2020). Using fractional derivatives for improved viscoelastic modeling of textile composites. Part II: Fabric under different temperatures. Composite Structures, 248, 112494. https://doi.org/10.1016/j.compstruct.2020.112494
Ferguson, M., Ak, R., Lee, Y. T. T., & Law, K. H. (2018). Detection and segmentation of manufacturing defects with convolutional neural networks and transfer learning. Smart and Sustainable Manufacturing Systems, 2(1), 137–164. https://doi.org/10.1520/SSMS20180033
Girshick, R. (2015). Fast R-CNN. In Proceedings of the IEEE international conference on computer vision, 2015 Inter (pp. 1440–1448). https://doi.org/10.1109/ICCV.2015.169
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014). Rich feature hierarchies for accurate object detection and semantic segmentation. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. https://doi.org/10.1109/CVPR.2014.81
Gupta, K., Körber, M., Djavadifar, A., Krebs, F., & Najjaran, H. (2019). Wrinkle and boundary detection of fiber products in robotic composites manufacturing. Assembly Automation, 40(2), 283–291. https://doi.org/10.1108/AA-05-2019-0094
Gupta, K, Körber, M., Krebs, F., & Najjaran, H. (2018). Vision-based deformation and wrinkle detection for semi-finished fiber products on curved surfaces. In 2018 IEEE 14th international conference on automation science and engineering (CASE) (pp. 618–623). https://doi.org/10.1109/COASE.2018.8560559
Habboush, A., Sanbhal, N., Shao, H., Jiang, J., & Chen, N. (2018). Characterization and analysis of in-plane shear behavior of glass warp-knitted non-crimp fabrics based on picture frame method. Materials. https://doi.org/10.3390/ma11091550
He, K., Gkioxari, G., Dollar, P., & Girshick, R. (2017). Mask R-CNN. In Proceedings of the IEEE international conference on computer vision, 2017-Octob (pp. 2980–2988). https://doi.org/10.1109/ICCV.2017.322
Huang, H., Wei, Z., & Yao, L. (2019). A novel approach to component assembly inspection based on mask R-CNN and support vector machines. Information (Switzerland). https://doi.org/10.3390/info10090282
Huang, J., Boisse, P., Hamila, N., & Zhu, Y. (2020). Simulation of wrinkling during bending of composite reinforcement laminates. Materials, 13(10), 2374. https://doi.org/10.3390/ma13102374
Iwata, A., Inoue, T., Naouar, N., Boisse, P., & Lomov, S. V. (2018). Meso-macro simulation of the woven fabric local deformation in draping. In AIP conference proceedings (Vol. 1960, p. 020012). American Institute of Physics Inc. https://doi.org/10.1063/1.5034813
Karim, M. M., Doell, D., Lingard, R., Yin, Z., Leu, M. C., & Qin, R. (2019). A region-based deep learning algorithm for detecting and tracking objects in manufacturing plants. Procedia Manufacturing, 39, 168–177. https://doi.org/10.1016/j.promfg.2020.01.289
Körber, M., Manufacturing, C. F.-P., & 2019, undefined. (n.d.). Automated planning and optimization of a draping processes within the CATIA environment using a Python Software Tool. Elsevier. https://www.sciencedirect.com/science/article/pii/S2351978920301141. Accessed February 26, 2020.
Liang, B., Colmars, J., & Boisse, P. (2017). A shell formulation for fibrous reinforcement forming simulations. Composites Part A: Applied Science and Manufacturing, 100, 81–96. https://doi.org/10.1016/j.compositesa.2017.04.024
Liang, B., Colmars, J., & Boisse, P. (2018). A shell approach for fibrous reinforcement forming simulations. In AIP conference proceedings (Vol. 1960, p. 020015). American Institute of Physics Inc. https://doi.org/10.1063/1.5034816
Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., et al. (2014). Microsoft coco: Common objects in context. In European conference on computer vision (pp. 740–755).
Long, J., Shelhamer, E., & Darrell, T. (2015). Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 3431–3440).
Mehdikhani, M., Petrov, N., Gorbatikh, L., & Lomov, S. V. (2018). Multi-scale experimental and computational investigation of matrix cracking evolution in carbon fiber-reinforced composites in the absence and presence of voids. In IOP conference series: Materials science and engineering (Vol. 406, p. 012011). Institute of Physics Publishing. https://doi.org/10.1088/1757-899X/406/1/012011
Montazerian, H., Sourki, R., Ramezankhani, M., Rashidi, A., Koerber, M., & Milani, A. S. (2019). Digital twining of an automated fabric draping process for industry 4.0 applications: Part imulti-body simulation and finite element modeling. In CAMX 2019: composites and advanced materials expo. https://elib.dlr.de/131180/. Accessed February 25, 2020.
Peng, X. Q., Cao, J., Chen, J., Xue, P., Lussier, D. S., & Liu, L. (2004). Experimental and numerical analysis on normalization of picture frame tests for composite materials. Composites Science and Technology, 64(1), 11–21. https://doi.org/10.1016/S0266-3538(03)00202-1
Poppe, C., Rosenkranz, T., Dörr, D., & Kärger, L. (2019). Comparative experimental and numerical analysis of bending behaviour of dry and low viscous infiltrated woven fabrics. Composites Part A: Applied Science and Manufacturing. https://doi.org/10.1016/j.compositesa.2019.05.034
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE computer society conference on computer vision and pattern recognition (Vol. 2016-Decem, pp. 779–788). https://doi.org/10.1109/CVPR.2016.91
Ren, S., He, K., Girshick, R., & Sun, J. (2015). Faster r-cnn: Towards real-time object detection with region proposal networks. In Advances in neural information processing systems (pp. 91–99).
Ronneberger, O., Fischer, P., & Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. In Lecture notes in computer science (including subseries lecture notes in artificial intelligence and lecture notes in bioinformatics) (Vol. 9351, pp. 234–241). Springer Verlag. https://doi.org/10.1007/978-3-319-24574-4_28
Singh, J., & Shekhar, S. (2018). Road damage detection and classification in smartphone captured images using mask R-CNN. In IEEE international conference on big data cup (Vol. abs/1811.0).
Tabernik, D., Šela, S., Skvarč, J., & Skočaj, D. (2020). Segmentation-based deep-learning approach for surface-defect detection. Journal of Intelligent Manufacturing, 31(3), 759–776. https://doi.org/10.1007/s10845-019-01476-x
Tang, T. W., Kuo, W. H., Lan, J. H., Ding, C. F., Hsu, H., & Young, H. T. (2020). Anomaly detection neural network with dual auto-encoders GAN and its industrial inspection applications. Sensors (Switzerland). https://doi.org/10.3390/s20123336
Voggenreiter, H., & Nieberl, D. (2015). AZIMUT Abschlussbericht. TIB.
Wang, D., Naouar, N., Vidal-Salle, E., & Boisse, P. (2018). Longitudinal compression and Poisson ratio of fiber yarns in meso-scale finite element modeling of composite reinforcements. Composites Part B: Engineering, 141, 9–19. https://doi.org/10.1016/j.compositesb.2017.12.042
Wang, K. J., Rizqi, D. A., & Nguyen, H. P. (2020). Skill transfer support model based on deep learning. Journal of Intelligent Manufacturing. https://doi.org/10.1007/s10845-020-01606-w
Wei, W., Zhang, C., & Deng, D. (2020). Content estimation of foreign fibers in cotton based on deep learning. Electronics (Switzerland), 9(11), 1–22. https://doi.org/10.3390/electronics9111795
Wen, V., Wong, H., Ferguson, M., Law, K. H., Lee, Y.-T. T., & Witherell, P. (2020). Automatic volumetric segmentation of additive manufacturing defects with 3D U-Net. SSS20 (VIRTUAL) AAAI spring symposium on AI in manufacturing. https://www.nist.gov/publications/automatic-volumetric-segmentation-additive-manufacturing-defects-3d-u-net
Xue, P., Peng, X., & Cao, J. (2003). A non-orthogonal constitutive model for characterizing woven composites. Composites Part A: Applied Science and Manufacturing, 34(2), 183–193. https://doi.org/10.1016/S1359-835X(02)00052-0
Yan, S., Zeng, X., & Long, A. (2019). Meso-scale modelling of 3D woven composite T-joints with weave variations. Composites Science and Technology, 171, 171–179. https://doi.org/10.1016/j.compscitech.2018.12.024
Zhao, G., Hu, J., Xiao, W., & Zou, J. (2020). A mask R-CNN based method for inspecting cable brackets in aircraft. Chinese Journal of Aeronautics. https://doi.org/10.1016/j.cja.2020.09.024
Zhao, H., Qi, X., Shen, X., Shi, J., & Jia, J. (2018). ICNet for real-time semantic segmentation on high-resolution images. In Lecture notes in computer science (including subseries lecture notes in artificial intelligence and lecture notes in bioinformatics), 11207 LNCS (pp. 418–434). https://doi.org/10.1007/978-3-030-01219-9_25
Acknowledgements
The authors would like to acknowledge the support of the Natural Sciences and Engineering Council Canada (NSERC) and Kinova Robotics under the NSERC Collaborative Research and Development (CRD) program Grant CRDPJ 543881-19, and the German Aerospace Center (DLR) towards this research.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations
Rights and permissions
About this article
Cite this article
Djavadifar, A., Graham-Knight, J.B., Kӧrber, M. et al. Automated visual detection of geometrical defects in composite manufacturing processes using deep convolutional neural networks. J Intell Manuf 33, 2257–2275 (2022). https://doi.org/10.1007/s10845-021-01776-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10845-021-01776-1