基于深度学习与图像处理技术的混凝土表面裂缝智能识别
Intelligent Recognition of Concrete Surface Cracks Based on Deep Learning and Image Processing Technology
摘要: 裂缝检测是保证各类设施安全的关键任务。然而，由于复杂的结构表面噪声与干扰阻碍裂缝的识别，深度学习技术识别裂缝后，无法进一步获得裂缝的相关参数，裂缝检测仍然具有一定挑战性。为解决上述问题，提出了基于深度学习与图像处理技术相结合的混凝土表面裂缝智能识别方法。在U-Net基础上，采用多级特征融合与引入图像梯度构建网络，实现裂缝的提取。构建包含连通域去噪、断裂连接、边缘检测、裂缝骨架化、裂缝参数计算等流程的图像处理技术，实现裂缝参数信息自动获取。通过实验验证，改进的U-net算法提取裂缝，其mPA达到90.08%。图像处理技术计算裂缝参数，裂缝宽度与实际值相接近，相对误差小于10%。研究结果表明：本文算法裂缝检测结果精度较高，且裂缝参数计算精度达到工程应用要求。
Abstract: Crack detection is a critical task to ensure the safety of all types of facilities. However, the complex structural surface noise and interference hinder the identification of cracks. After the deep learning technology identifies the crack, the relevant parameters of the crack cannot be further obtained, and the crack detection is still challenging. In order to solve the above problems, an intelligent identification method for concrete surface cracks based on the combination of deep learning and image processing technology is proposed. On the basis of U-Net, the network is constructed by mul-ti-level feature fusion and the introduction of image gradient to realize the extraction of cracks, and the image processing technology including connected domain denoising, fracture connection, edge detection, fracture skeletonization, fracture parameter calculation and other processes is con-structed to realize the automatic acquisition of fracture parameter information. Through experi-mental verification, the improved U-net algorithm can extract cracks and make the mPA reach 90.08%. The image processing technology calculates the crack parameters, and the crack width is close to the actual value, and the relative error is less than 10%. The research results show that the algorithm in this paper has high accuracy of crack detection results, and the calculation accuracy of crack parameters meets the requirements of engineering application.
文章引用：隆涛. 基于深度学习与图像处理技术的混凝土表面裂缝智能识别[J]. 应用数学进展, 2023, 12(3): 1130-1140. https://doi.org/10.12677/AAM.2023.123115

参考文献

[1] Zakeri, H., Nejad, F.M. and Fahimifar, A. (2017) Image Based Techniques for Crack Detection, Classification and Quan-tification in Asphalt Pavement: A Review. Archives of Computational Methods in Engineering, 24, 935-977. [Google Scholar] [CrossRef] 
[2] Sattar, D., Thomas, R.J. and Marc, M. (2018) Comparison of Deep Convolutional Neural Networks and Edge Detectors for Image-Based Crack Detection in Concrete. Construction and Building Materials, 186, 1031-1045. [Google Scholar] [CrossRef] 
[3] Hsieh, Y.A. and Tsai, Y.J. (2020) Machine Learning for Crack Detection: Review and Model Performance Comparison. Journal of Computing in Civil Engineering, 34, Article ID: 04020038. [Google Scholar] [CrossRef] 
[4] Cha, Y.J., Choi, W. and Buyukozturk, O. (2017) Deep Learning-Based Crack Damage Detection Using Convolutional Neural Networks. Computer-Aided Civil and Infra-structure Engineering, 32, 361-378. [Google Scholar] [CrossRef] 
[5] Ren, S., He, K., Girshick, R., et al. (2017) Faster R-CNN: Towards Re-al-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis & Machine Intelli-gence, 39, 1137-1149. [Google Scholar] [CrossRef] 
[6] Zhang, Y., Huang, J. and Cai, F. (2020) On Bridge Surface Crack Detection Based on an Improved YOLO v3 Algorithm. IFAC-PapersOnLine, 53, 8205-8210. [Google Scholar] [CrossRef] 
[7] Lecun, Y., Bengio, Y. and Hinton, G. (2015) Deep Learning. Nature, 521, 436-444. [Google Scholar] [CrossRef] [PubMed]
[8] Badrinarayanan, V., Kendall, A. and Cipolla, R. (2017) SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation. IEEE Transactions on Pattern Analysis and Ma-chine Intelligence, 39, 2481-2495. [Google Scholar] [CrossRef] 
[9] Zhao, H., Shi, J., Qi, X., et al. (2017) Pyramid Scene Parsing Network. 2017 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, 25-27 April 2017, 6230-6239. [Google Scholar] [CrossRef] 
[10] Chen, L.C., Papandreou, G., Kokkinos, I., et al. (2017) Deeplab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected Crfs. IEEE Transactions on Pattern Analysis and Machine Intelligence, 40, 834-848. [Google Scholar] [CrossRef] 
[11] Ronneberger, O., Fischer, P. and Brox, T. (2015) U-Net: Convolutional Networks for Biomedical Image Segmentation. 18th International Conference, Munich, 5-9 October 2015, 234-241. [Google Scholar] [CrossRef] 
[12] Ren, Y., Huang, J., Hong, Z., et al. (2020) Im-age-Based Concrete Crack Detection in Tunnels Using Deep Fully Convolutional Networks. Construction and Building Materials, 234, 117367. [Google Scholar] [CrossRef] 
[13] 于海洋, 景鹏, 张文涛, 等. 基于残差和注意力机制的道路裂缝检测U-Net改进模型[J]. 计算机工程, 2022, 1-14.
[14] 李红卫, 熊韬. 基于二维码识别的P4P算法研究[J]. 激光与光电子学进展, 2020, 57(12): 304-312.
[15] Ren, X., Xing, Z., Xia, X., Grundy, et al. (2019) Neural Net-work-Based Detection of Self-Admitted Technical Debt: From Performance to Explainability. ACM Transactions on Software Engineering and Methodology, 28, 15.1-15.45. [Google Scholar] [CrossRef] 
[16] Eelbode, T., Bertels, J., Berman, M., et al. (2020) Optimization for Medical Image Segmentation: Theory and Practice When Evaluating with Dice Score or Jaccard Index. IEEE Transactions on Medical Imaging, 39, 3679-3690. [Google Scholar] [CrossRef] 
[17] 陈瑶. 基于图像分析的桥梁裂缝检测方法研究[D]: [硕士学位论文]. 合肥: 中国科学技术大学, 2016.
[18] Zhang, T.Y. and Suen, C.Y. (1984) A Fast Parallel Algorithm for Thin-ning Digital Patterns. Communications of the ACM, 27, 236-239. [Google Scholar] [CrossRef] 
[19] Grelard, F., Baldacci, F., Vialard, A., et al. (2017) New Methods for the Geometrical Analysis of Tubular Organs. Medical Image Analysis, 42, 89-101. [Google Scholar] [CrossRef] [PubMed]
[20] Fu, H.X., Meng, D., Li, W.H., et al. (2021) Bridge Crack Seman-tic Segmentation Based on Improved Deeplabv3+. Journal of Marine Science and Engineering, 9, 671. [Google Scholar] [CrossRef] 

为你推荐