|
|
改进YOLOv7的地下工程衬砌表观病害检测算法
|
Abstract:
针对地下工程衬砌表观病害检测精度低的问题，给出了一种改进YOLOv7的衬砌表观病害检测算法。该算法首先在主干网络输出位置各自加入CBAM注意力机制，用于加强主干网络对关键区域的特征提取能力；其次将无参数注意力机制SimAM引入到SPPCSPC模块中，并裁剪掉冗余的一层CBS模块，在增强主干网络对密集小目标的特征提取能力的同时，缩减计算量和参数规模；最后使用自制数据集训练算法模型，与原算法相比本算法的P、R以及mAP0.5分别提升了12.7、5.4和8.6个百分点。实验结果表明本文改进方法有效地提升了算法对衬砌表观病害的检测性能。
Aiming at the low detection accuracy of underground engineering lining apparent disease, an improved YOLOv7 algorithm for lining apparent disease detection is presented. Firstly, CBAM attention mechanism is added to each output position of the backbone network to enhance the feature extraction capability of the backbone network for key regions. Secondly, the non-parametric attention mechanism SimAM is introduced into SPPCSPC module, and the redundant CBS module is cut out, which enhances the feature extraction ability of the backbone network for dense small targets, and reduces the computation and parameter scale. Finally, the algorithm model is trained with self-made data set. Compared with the original algorithm, the P, R and mAP0.5 of the proposed algorithm are improved by 12.7, 5.4 and 8.6 percentage points respectively. The results show that the improved method can effectively improve the detection performance of the algorithm on the apparent diseases of the lining.
| [1] | 陈湘生, 付艳斌, 陈曦, 等. 地下空间施工技术进展及数智化技术现状[J]. 中国公路学报, 2022, 35(1): 1-12. https://doi.org/10.19721/j.cnki.1001-7372.2022.01.001 |
| [2] | 刘德军, 仲飞, 黄宏伟, 等. 运营隧道衬砌病害诊治的现状与发展[J]. 中国公路学报, 2021, 34(11): 178-199. https://doi.org/10.19721/j.cnki.1001-7372.2021.11.015 |
| [3] | 《中国公路学报》编辑部. 中国隧道工程学术研究综述·2015 [J]. 中国公路学报, 2015, 28(5): 1-65. |
| [4] | 范文田. 我国隧道及地下工程的现状和发展[J]. 土木工程学报, 1995(6): 73-75. |
| [5] | 王平让. 隧道病害快速检测技术现状及分析比较[J]. 公路与汽运, 2016(3): 241-245. |
| [6] | 卢芳芳. 基于深度学习的隧道衬砌病害检测与识别研究[D]: [硕士学位论文]. 郑州: 华北水利水电大学, 2022. https://doi.org/10.27144/d.cnki.ghbsc.2021.000314 |
| [7] | 任松, 朱倩雯, 涂歆玥, 等. 基于深度学习的公路隧道衬砌病害识别方法[J]. 浙江大学学报(工学版), 2022, 56(1): 92-99. |
| [8] | 胡利娜. 一种基于深度学习的隧道衬砌病害检测技术[J]. 山西电子技术, 2019(5): 38-40. |
| [9] | 任安虎, 姜子渊, 马晨浩. 基于改进YOLOv5s的道路裂缝检测算法[J/OL]. 激光杂志: 1-7. http://kns.cnki.net/kcms/detail/50.1085.TN.20230626.1756.012.html, 2024-03-18. |
| [10] | Wang, C.Y., Bochkovskiy, A. and Liao, H.Y.M. (2023) YOLOv7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors. Computer Vision and Pattern Recognition. https://doi.org/10.1109/CVPR52729.2023.00721 |
| [11] | Ding, X., Zhang, X., Ma, N., et al. (2021) Repvgg: Making Vgg-Style Convnets Great Again. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Nashville, TN, 20-25 June 2021, 13733-13742. https://doi.org/10.1109/CVPR46437.2021.01352 |
| [12] | Woo, S., Park, J., Lee, J.Y., et al. (2018) Cbam: Convolutional Block Attention Module. Proceedings of the European Conference on Computer Vision (ECCV), 3-19. https://doi.org/10.1007/978-3-030-01234-2_1 |
| [13] | Yang, L., Zhang, R.Y., Li, L., et al. (2021) Simam: A Simple, Parameter-Free Attention Module for Convolutional Neural Networks. International Conference on Machine Learning, 11863-11874. |
| [14] | Dadboud, F., Patel, V., Mehta, V., et al. (2021) Singlestage UAV Detection and Classification with YOLOV5: Mosaic Data Augmentation and PANet. 2021 17th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), Washington DC, 16-19 November 2021, 1-8. https://doi.org/10.1109/AVSS52988.2021.9663841 |
