基于改进Mask R-CNN深度学习算法的隧道裂缝智能检测方法

doi:10.11996/JG.j.2095-302X.2023010177

图学学报 ›› 2023, Vol. 44 ›› Issue (1): 177-183.DOI: 10.11996/JG.j.2095-302X.2023010177

基于改进Mask R-CNN深度学习算法的隧道裂缝智能检测方法

朱磊¹(), 李东彪¹, 闫星志², 刘向阳², 沈才华³()

1.中交隧桥(南京)技术有限公司，江苏南京 211800
2.河海大学理学院，江苏南京 211100
3.河海大学土木与交通学院，江苏南京 210098

收稿日期:2022-05-09 修回日期:2022-07-17 出版日期:2023-10-31 发布日期:2023-02-16
通讯作者: 沈才华
作者简介:朱磊(1983-)，男，硕士研究生。主要研究方向为道路与隧道养护技术、材料及装备研发。E-mail：361937742@qq.com
基金资助:
国家自然科学基金项目(41830110)

Intelligent detection method of tunnel cracks based on improved Mask R-CNN deep learning algorithm

ZHU Lei¹(), LI Dong-biao¹, YAN Xing-zhi², LIU Xiang-yang², SHEN Cai-hua³()

1. China Communications Construction Company Tunnel and Bridge (Nanjing) Technology Co., Ltd, Nanjing Jiangsu 211800, China
2. College of Science, Hohai University, Nanjing Jiangsu 211100, China
3. College of Civil and Transportation Engineering, Hohai University, Nanjing Jiangsu 210098, China

Received:2022-05-09 Revised:2022-07-17 Online:2023-10-31 Published:2023-02-16
Contact: SHEN Cai-hua
About author:ZHU Lei (1983-), master student. His main research interests cover tunnel equipment and road maintenance technology. E-mail：361937742@qq.com
Supported by:
National Natural Science Foundation of China(41830110)

摘要/Abstract

摘要：

隧道裂缝检测是避免隧道重大灾害和日常养护的重要工作，但传统人工检测工作量巨大，无法满足实际需求。采用深度学习神经网络Mask R-CNN模型对裂缝进行智能自动检测，避免了人工检测的耗时耗力。通过调整算法参数，优化模型检测结果，获得适用于隧道裂缝检测的Mask R-CNN模型。针对自动识别的裂缝结果，进一步计算其几何特征参数。为充分利用裂缝狭长弯曲特性，体现裂缝走向及基本形态，提出了基于骨架提取和函数拟合思想的裂缝几何特征计算方法。根据裂缝骨架，可获得裂缝走势，计算裂缝长度。通过函数拟合，可得到贯穿裂缝狭长区域的函数，根据函数法向量计算宽度。根据裂缝几何参数计算结果，结合规范规定的须修复裂纹宽度要求，可实现隧道裂缝检测自动预警，为隧道裂缝自动检测提供了技术支撑。

关键词: 隧道裂缝, 深度学习, Mask R-CNN, 骨架提取

Abstract:

Tunnel crack detection is an important task for the prevention of major disasters and daily maintenance of the tunnel, but traditional manual detection would incur huge workload and cannot meet practical needs. The deep learning neural network Mask R-CNN model was used for intelligent automatic detection of cracks, avoiding the time-consuming and labor-intensive manual detection. By adjusting the algorithm parameters and optimizing the model detection results, the Mask R-CNN model suitable for tunnel crack detection was obtained. For the results of automatic identification of cracks, the geometric characteristic parameters were further calculated. In order to make full use of the long and narrow bending characteristics of cracks and reflect the trend and basic shape of cracks, a calculation method of geometric characteristics of cracks based on skeleton extraction and function fitting was proposed. According to the crack skeleton, the crack trend could be obtained and the crack length could be calculated. Through function fitting, the function penetrating the narrow and long region of the crack could be obtained, and the width could be calculated according to the normal vector of the function. According to the calculation results of crack geometric parameters, combined with the requirements of crack width to be repaired specified in the specification, the automatic early warning of tunnel crack detection could be realized, providing technical support for the automatic detection of tunnel cracks.

Key words: tunnel crack, deep learning, Mask R-CNN, skeleton extraction

中图分类号:

朱磊, 李东彪, 闫星志, 刘向阳, 沈才华. 基于改进Mask R-CNN深度学习算法的隧道裂缝智能检测方法[J]. 图学学报, 2023, 44(1): 177-183.

ZHU Lei, LI Dong-biao, YAN Xing-zhi, LIU Xiang-yang, SHEN Cai-hua. Intelligent detection method of tunnel cracks based on improved Mask R-CNN deep learning algorithm[J]. Journal of Graphics, 2023, 44(1): 177-183.

图/表 11

图1 裂缝数据集标注((a)裂缝区域选定；(b)裂缝掩膜)

Fig. 1 Crack dataset annotation ((a) Crack annotation; (b) Crack mask)

图2 Mask R-CNN网络结构图

Fig. 2 Mask R-CNN network structure diagram

图3 裂缝的几何参数提取算法流程图

Fig. 3 Flow chart of fracture geometric parameter extraction algorithm

图4 裂缝图像宽度计算示意图

Fig. 4 Schematic diagram of crack image width

表1 Mask R-CNN算法参数

Table 1 Mask R-CNN Algorithm parameters

参数设置	相关说明
Batch size=2	批处理个数，此处为2
Epochs=200	训练200轮数据集
Learning_rate=0.004	学习率设置为0.004
Weight decay=0.000 1	权重衰减系数为0.000 1

图5 Mask R-CNN模型训练损失值曲线

Fig. 5 Mask R-CNN model training loss curve

表2 混淆矩阵

Table 2 Confusion matrix

混淆矩阵		真实值
混淆矩阵		Positive	Negative
预测值	Positive	真正例(TP)	假正例(FP)
预测值	Negative	假反例(FN)	真反例(TN)

图6 Mask R-CNN模型训练过程中mAP曲线

Fig. 6 Mask R-CNN model training mAP curve

图7 部分检测结果((a)原始图像；(b) Mask R-CNN分割结果；(c)全局阈值二值化分割结果)

Fig. 7 Partial detection results ((a) Original image; (b) Mask R-CNN segmentation results; (c) Global threshold binarization segmentation results)

图8 部分裂缝掩膜骨架提取效果展示((a)原始图像；(b)裂缝掩膜；(c)骨架提取结果)

Fig. 8 Crack skeletonize results ((a) Original image; (b) Crack mask; (c) Skeleton extraction results)

图9 部分裂缝骨架函数拟合效果((a)函数拟合效果1；(b)函数拟合效果2)

Fig. 9 Function fitting results of crack skeleton ((a) Function fitting effect 1; (b) Function fitting effect 2)

参考文献 11

[1]	中华人民共和国交通运输部. 2021年交通运输行业发展统计公报[R/OL]. (2022-05-28) [2022-07-12].http://www.gov.cn/xinwen/2022-05/25/content_5692174.htm.
	Ministry of Transport of the Peopleʹs Republic of China. Statistical bulletin of transportation industry development in 2021[R/OL]. (2022-05-28) [2022-07-12].http://www.gov.cn/xinwen/2022-05/25/content_5692174.htm.
[2]	ZHANG L, YANG F, ZHANG Y D, et al. Road crack detection using deep convolutional neural network[C]//IEEE International Conference on Image Processing. New York: IEEE Press, 2016: 3708-3712.
[3]	CHENG J, XIONG W, CHEN W, et al. Pixel-level crack detection using U-net[C]//2018 IEEE Region 10 Conference. New York: IEEE Press, 2018: 462-466.
[4]	FAN Z, LI C, CHEN Y, et al. Automatic crack detection on road pavements using encoder-decoder architecture[J]. Materials, 2020, 13(13): E2960.
[5]	李想, 熊进刚. 基于深度学习的低光照条件的混凝土裂缝检测[J]. 建筑结构, 2021, 51(S2): 1046-1050.
	LI X, XIONG J G. Detection of concrete cracks in low light condition based on deep learning[J]. Building Structure, 2021, 51(S2): 1046-1050 (in Chinese).
[6]	张世瑶, 贺玉彬, 周新志. 基于深度学习的道路表面裂缝检测方法[J]. 科学技术与工程, 2021, 21(15): 6380-6385.
	ZHANG S Y, HE Y B, ZHOU X Z. Road surface crack detection method based on deep learning[J]. Science Technology and Engineering, 2021, 21(15): 6380-6385 (in Chinese).
[7]	李良福, 马卫飞, 李丽, 等. 基于深度学习的桥梁裂缝检测算法研究[J]. 自动化学报, 2019, 45(9): 1727-1742.
	LI L F, MA W F, LI L, et al. Research on detection algorithm for bridge cracks based on deep learning[J]. Acta Automatica Sinica, 2019, 45(9): 1727-1742 (in Chinese).
[8]	蔡光明. 高速铁路桥梁底面裂缝的视频检测系统研究[D]. 北京: 北京交通大学, 2011.
	CAI G M. Study on the video detection system of crack on the bottom of high-speed rail bridge[D]. Beijing: Beijing Jiaotong University, 2011 (in Chinese).
[9]	高新闻, 李帅青, 金邦洋. 基于DenseNet分类的隧道裂缝检测研究[J]. 计算机测量与控制, 2020, 28(8): 58-61, 87.
	GAO X W, LI S Q, JIN B Y. Study on tunnel crack detection based on DenseNet classification[J]. Computer Measurement & Control, 2020, 28(8): 58-61, 87 (in Chinese).
[10]	HE K, GKIOXARI G, PIOTR DOLLAR, et al. Mask R-CNN[C]//2017 IEEE International Conference on Computer Vision. New York: IEEE Press, 2017: 2961-2969.
[11]	REN S, HE K, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149. DOI PMID

[1]	杨陈成, 董秀成, 侯兵, 张党成, 向贤明, 冯琪茗. 基于参考的Transformer纹理迁移深度图像超分辨率重建[J]. 图学学报, 2023, 44(5): 861-867.
[2]	党宏社, 许怀彪, 张选德. 融合结构信息的深度学习立体匹配算法[J]. 图学学报, 2023, 44(5): 899-906.
[3]	翟永杰, 郭聪彬, 王乾铭, 赵宽, 白云山, 张冀. 基于隐含空间知识融合的输电线路多金具检测方法[J]. 图学学报, 2023, 44(5): 918-927.
[4]	杨红菊, 高敏, 张常有, 薄文, 武文佳, 曹付元. 一种面向图像修复的局部优化生成模型[J]. 图学学报, 2023, 44(5): 955-965.
[5]	毕春艳, 刘越. 基于深度学习的视频人体动作识别综述[J]. 图学学报, 2023, 44(4): 625-639.
[6]	曹义亲, 周一纬, 徐露. 基于E-YOLOX的实时金属表面缺陷检测算法[J]. 图学学报, 2023, 44(4): 677-690.
[7]	邵俊棋, 钱文华, 徐启豪. 基于条件残差生成对抗网络的风景图生成[J]. 图学学报, 2023, 44(4): 710-717.
[8]	余伟群, 刘佳涛, 张亚萍. 融合注意力的拉普拉斯金字塔单目深度估计[J]. 图学学报, 2023, 44(4): 728-738.
[9]	郭印宏, 王立春, 李爽. 基于重复性和特异性约束的图像特征匹配[J]. 图学学报, 2023, 44(4): 739-746.
[10]	毛爱坤, 刘昕明, 陈文壮, 宋绍楼. 改进YOLOv5算法的变电站仪表目标检测方法[J]. 图学学报, 2023, 44(3): 448-455.
[11]	王佳婧, 王晨, 朱媛媛, 王笑梅. 基于民国纸币的图元素匹配检索[J]. 图学学报, 2023, 44(3): 492-501.
[12]	杨柳, 吴晓群. 基于深度学习的三维形状补全研究综述[J]. 图学学报, 2023, 44(2): 201-215.
[13]	曾武, 朱恒亮, 邢树礼, 林江宏, 毛国君. 显著性检测引导的图像数据增强方法[J]. 图学学报, 2023, 44(2): 260-270.
[14]	罗启明, 吴昊, 夏信, 袁国武. 基于Dual Dense U-Net的云南壁画破损区域预测[J]. 图学学报, 2023, 44(2): 304-312.
[15]	李洪安, 郑峭雪, 陶若霖, 张敏, 李占利, 康宝生. 基于深度学习的图像超分辨率研究综述[J]. 图学学报, 2023, 44(1): 1-15.

基于改进Mask R-CNN深度学习算法的隧道裂缝智能检测方法

Intelligent detection method of tunnel cracks based on improved Mask R-CNN deep learning algorithm

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 11

相关文章 15

编辑推荐

Metrics

本文评价