胡海燕,张娟娟,宋圭辰,李硕,刘昱萌,刘斌
HU Hai-yan,ZHANG Juan-juan,SONG Gui-chen,LI Shuo,LIU Yu-men,LIU Bin

• 1:国网陕西省电力有限公司榆林供电公司
• 2:陕西科技大学电子信息与人工智能学院
• 3:宁波大学材料科学与化学工程学院

摘要(Abstract)：

扫描电子显微镜(SEM)可以有效地观测到碳纤维复合材料(CFRP)中碳纤维(CF)的形态和分布，但是对CF进行定性观察对改善CFPR电性能的贡献是有限的。在Mask R-CNN的基础上提出了SoftMask R-CNN来实现CF的SEM图像自动分割，进行CF分布的评估，研究CF分布对CFRP电性能的影响。试验结果表明：Soft-Mask R-CNN在SEM图像上的平均准确率和交并比分别为86.9%、90.7%;Soft-Mask R-CNN在不同的SEM放大条件下具有稳定的分割结果；Soft-Mask R-CNN对CF的SEM图像进行实时分割满足了对连续SEM图像观测的需求，表明CF分布可以改善CFRP的电性能。
Scanning electron microscope(SEM) can effectively observe the morphology and distribution of carbon fibers(CF) in carbon fiber reinforced composite materials(CFRP), but the contribution of qualitative observation of CF to improving the electrical properties of CFPR is limited. Therefore, on the basis of Mask R-CNN, Soft-Mask R-CNN is proposed to realize automatic segmentation of CF SEM images, evaluate CF distribution, and study the influence of CF distribution on the electrical performance of CFRP. Experimental results show that the average accuracy and intersection ratio of Soft-Mask R-CNN on SEM images are86.9% and 90.7% respectively; Soft-Mask R-CNN has stable segmentation results under different SEM magnification conditions; Soft-Mask R-CNN's real-time segmentation of CF SEM images meets the demand for continuous SEM image observation, indicating that CF distribution can improve the electrical performance of CFRP.

关键词(KeyWords)： 碳纤维复合材料;碳纤维分布;Mask R-CNN
carbon fiber reinforced composite material;carbon fiber distribution;Mask R-CNN

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金(61871260)

作者(Author): 胡海燕,张娟娟,宋圭辰,李硕,刘昱萌,刘斌
HU Hai-yan,ZHANG Juan-juan,SONG Gui-chen,LI Shuo,LIU Yu-men,LIU Bin

DOI: 10.16090/j.cnki.hcxw.2022.11.024

参考文献(References)：

• [1] DOBRZASKI L A, PUSZ A, NOWAK A J. Aramid-silicon laminated materials with special properties-new perspective of its usage[J]. Journal of Achievements in Materials and Manufacturing Engineering, 2008, 28(1):7-14.
• [2] HUFENBACH W, DOBRZANSKI L A, GUDE M, et al. Optimisation of the rivet joints of the CFRP composite material and aluminium alloy[J].Journal of Achievements in Materials and Manufacturing Engineering,2007, 20(1-2):24-28.
• [3] CHIARELLO M, ZINNO R. Electrical conductivity of self-monitoring CFRP[J]. Cement and Concrete Composites, 2005, 27(4):463-469.
• [4] RAFIQUE I, KAUSAR A, MUHAMMAD B. Epoxy resin composite reinforced with carbon fiber and inorganic filler:overview on preparation and properties[J]. Polymer-Plastics Technology and Engineering, 2016, 55(15):1653-1672.
• [5] GAO J, GUO H Y, WANG X F, et al. Microwave deicing for asphalt mixture containing steel wool fibers[J]. Journal of Cleaner Production,2019, 206:1110-1122.
• [6] WANG C, JIAO G S, LI B L, et al. Dispersion of carbon fibers and conductivity of carbon fiber-reinforced cement-based composites[J]. Ceramics International, 2017, 43(17):15122-15132.
• [7] LU Y H, WANG J, XU R X. et al. Effects of different carbon fiber content on conductive composites'electrical proper-ties[C]//The 7th National Conference on Functional Materials and Applications, 2010:2218-2220.
• [8] LONG J, SHELHAMER E, DARRELL T. Fully convolutional networks for semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition(CVPR), 2015:3431-3440.
• [9] RONNEBERGER O, FISCHER P, BROX T. U-Net:convolutional networks for biomedical image segmentation[C]//International Conference on Medical Image Computing and Computerassisted Intervention(MICCAI),2015:234-241.
• [10] TONG Z, GUO H Y, GAO J, et al. A novel method for multi-scale carbon fiber distribution characterization in cement-based composites[J].Construction and Building Materials, 2019, 218:40-52.
• [11] 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(Basel),2020, 13(4):936.
• [12]王雯,张芳.基于深度学习的SEM纤维图像分割方法研究[J].中国纤检, 2019, 52(2):84-87.
• [13] PONIKIEWSKI T, KATZER J, BUGDOL M, et al. Determination of3D porosity in steel fibre reinforced SCC beams using X-ray computed tomography[J]. Construction and Building Materials, 2014, 68:333-340.
• [14] PONIKIEWSKI T, KATZER J, BUGDOL M, et al. Steel fibre spacing in self-compacting concrete precast walls by X-ray computed tomography[J]. Materials and Structures, 2015, 48(12):3863-3874.
• [15] HE K M, GKIOXARI G, DOLLAR P, et al. Mask R-CNN[J]. IEEE Transactions on Pattern Analysis&Machine Intelligence, 2017, 42(2):386-397.
• [16] REN S, HE K M, GIRSHICK R, et al. Faster R-CNN:towards realtime object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis&Machine Intelligence, 2017, 39(6):1137-1149.
• [17] HU J, LI S, ALBANIE S, et al. Squeeze-and-excitation networks[J].IEEE Computer Society, 2020, 42(8):2011-2023.
• [18] NEUBECK A, VAN G L. Efficient non-maximum suppression[C]//18th International Conference on Pattern Recognition, 2006:9210072.
• [19] CHUNG D D, Dispersion of short fibers in cement[J]. Journal of Materials in Civil Engineering, 2005, 17(4):379-383.