基于掌子面图像识别的隧道围岩分级与应用

徐敏; 余望; 傅鹤林; 曹桂乾; 李俊

doi:10.20265/j.cnki.issn.1007-2993.2025-0053

基于掌子面图像识别的隧道围岩分级与应用

doi: 10.20265/j.cnki.issn.1007-2993.2025-0053

徐敏^1,,
余望²,
傅鹤林^3, ,,
曹桂乾^3,,
李俊²

1.
岳阳市城市运营投资集团有限公司，湖南岳阳　414000
2.
五矿二十三冶建设集团有限公司，湖南长沙　410116
3.
中南大学，湖南长沙　410075

基金项目: 湖南省教育厅重点课题（23A0014）

详细信息

作者简介:
徐　敏，男，1988年生，大学本科，高级工程师，主要从事市政和建筑工程管理工作。E-mail：512306774@qq.com

通讯作者:
傅鹤林，男，1965年生，博士，二级教授，主要从事隧道工程灾害防治研究。E-mail：fu.h.l@csu.edu.cn

中图分类号: U45
计量
- 文章访问数: 9
- HTML全文浏览量: 2
- PDF下载量: 0
- 被引次数: 0
出版历程
- 收稿日期: 2025-02-10
- 修回日期: 2025-04-18
- 录用日期: 2025-06-26
- 网络出版日期: 2026-04-09
- 刊出日期: 2026-04-09

Tunnel surrounding rock classification and application based on face image recognition

XU Min^1
,,
YU Wang²,
FU Helin^{3
, ,},
CAO Guiqian^3
,,
LI Jun²

1.
Yueyang City Operation Investment Group Co., Ltd., Yueyang 414000, Hunan, China
2.
The 23^rd Metallurgical Construction Group Co., Ltd. of Minmetals, Changsha 410116, Hunan, China
3.
Central South University, Changsha 410075, Hunan, China

摘要

摘要: 隧道开挖过程中围岩的稳定性极其重要，支护设计依赖于隧道围岩质量的精准分级。为实现隧道施工期间的围岩快速精准分级，基于YOLOv8深度学习算法和数字图像处理技术，以隧道掌子面为研究对象，结合修正BQ法提出了一种隧道围岩智能化快速分级方法。结果表明，YOLOv8深度学习模型能够准确识别并定位隧道掌子面照片上的裂隙位置，结合图像处理技术可有效提取掌子面裂隙信息，从而有效评价隧道掌子面完整程度，实现对隧道围岩的快速分级。该方法在岳阳来米坡隧道实际应用表明，与实际围岩等级相比，采用该方法的围岩等级预测准确率达90%，能够满足隧道施工期间围岩快速分级需求。研究成果为隧道围岩等级的动态划分及指导隧道开挖、支护提供了参考。
- 围岩快速分级 /
- 深度学习 /
- 图像识别 /
- YOLOv8算法
Abstract: The stability of surrounding rock during tunnel excavation is crucial, and support design relies on accurate classification of the surrounding rock quality. To achieve rapid and precise classification during tunnel construction, an intelligent and rapid classification method for tunnel surrounding rock was proposed, based on the YOLOv8 deep learning algorithm and digital image processing technology, focusing on the tunnel face and incorporating the modified BQ method. The results indicate that the YOLOv8 deep learning model can accurately identify and locate fractures in tunnel face photographs. Combining with image processing techniques, it effectively extracts fracture information from the face, thereby assessing the integrity of the tunnel face and enabling rapid classification of the surrounding rock. Field validation in Yueyang Laimipo Tunnel demonstrates that, compared to actual surrounding rock grades, this method achieves a prediction accuracy of 90%, meeting the needs for rapid classification during tunnel construction. The research provides a reference for dynamic classification of tunnel surrounding rock grades and offers guidance for tunnel excavation and support.
- rapid classification of surrounding rock /
- deep learning /
- image recognition /
- YOLOv8

HTML全文

图 1 YOLOv8算法检测结果

Figure 1. Detection results of YOLOv8 algorithm

下载: 全尺寸图片幻灯片

图 2 YOLOv8的PR曲线

Figure 2. PR curve of YOLOv8

下载: 全尺寸图片幻灯片

图 3 掌子面完整程度评价流程

Figure 3. Evaluation process of tunnel face integrity

下载: 全尺寸图片幻灯片

图 4 掌子面照片处理过程

Figure 4. Processing procedure of tunnel face photographs

下载: 全尺寸图片幻灯片

图 5 隧道掌子面回弹测试区及测点布置

Figure 5. Rebound test zones and measuring point arrangement on tunnel face

下载: 全尺寸图片幻灯片

图 6 围岩智能分级流程

Figure 6. Intelligent classification workflow for surrounding rock

下载: 全尺寸图片幻灯片

图 7 分级情况对比图

Figure 7. Comparison diagram of classification results

下载: 全尺寸图片幻灯片

表 1 BQ法围岩分级表^[24]

Table 1. Rock mass classification table using BQ method^[24]

岩体质量级别	岩体基本质量的定性特征	岩体质量指标
Ⅰ	坚硬岩，岩体完整	>550
Ⅱ	坚硬岩，岩体较完整；较坚硬岩，岩体完整	550～451
Ⅲ	坚硬岩，岩体较破碎；较坚硬岩，岩体较完整；较软岩，岩体完整	450～351
Ⅳ	坚硬岩，岩体破碎；较坚硬岩，岩体较破碎—破碎；较软岩，岩体较完整—较破碎；软岩，岩体完整—较完整	350～251
Ⅴ	较软岩，岩体破碎；软岩，岩体较破碎—破碎；全部极软岩及极破碎岩	≤250

下载: 导出CSV

表 2 地下水影响修正系数K₁^[24]

Table 2. Correction coefficient K₁ for groundwater influence^[24]

地下水状态	BQ值
地下水状态	>550	550～451	450～351	350～251	≤250
潮湿或点滴状出水，p≤0.1或Q≤25	0	0	0～0.1	0.2～0.3	0.4～0.6
淋雨状或线流状出水，0.1≤p≤0.5或25≤Q≤125	0～0.1	0.1～0.2	0.2～0.3	0.4～0.6	0.7～0.9
涌流状出水，p>0.5或Q>125	0.1～0.2	0.2～0.3	0.4～0.6	0.7～0.9	1.0
注：p为地下工程围岩裂隙水压，MPa；Q为每10 m洞长出水量，L/min·(10 m)。

下载: 导出CSV

表 3 软弱结构面产状影响修正系数K₂^[24]

Table 3. Correction coefficient K₂ for weak structural plane occurrence^[24]

结构面产状及其洞轴线的组合关系	结构面走向与洞轴线夹角<30°，结构面倾角30°～75°	结构面走向与洞轴线夹角>60°，结构面倾角>75°	其他组合关系
K₂	0.4～0.6	0～0.2	0.2～0.4

下载: 导出CSV

表 4 初始地应力状态修正系数K₃^[24]

Table 4. Correction coefficient K₃ for initial ground stress state^[24]

初始应力状态		BQ值
初始应力状态		>550	550～451	450～351	350～251	≤250
极高应力	R_c/σ_max<4	1.0	1.0	1.0～1.5	1.0～1.5	1.0
高应力	R_c/σ_max=4～7	0.5	0.5	0.5	0.5～1.0	0.5～1.0

下载: 导出CSV

表 5 裂隙比K_s划分表^[1]

Table 5. Classification table for fracture ratio K_s^[1]

裂隙比	>0.1	0.08～0.1	0.04～0.08	0.01～0.04	<0.01
节理裂隙发育程度	很发育	发育	一般发育	较不发育	不发育
掌子面完整程度	极破碎	破碎	较破碎	较完整	完整

下载: 导出CSV

表 6 实验超参数设置

Table 6. Experimental hyperparameter settings

参数名称	设置值
Image-size（图像尺寸）	640×640×3
epochs（迭代次数）	300
Batch size（批次大小）	4
optimizer（优化器）	SGD
Learing rate（学习率）	0.01
Momentum（动量）	0.937
Weight decay（权重衰减）	0.0005

下载: 导出CSV

表 7 裂隙比K_s和完整性系数K_v对应关系表

Table 7. Corresponding relationship between fracture ratio K_s and integrity coefficient K_v

掌子面完整程度	极破碎	破碎	破碎	较完整	完整
裂隙比K_s	>0.1	0.08～0.1	0.04～0.08	0.01～0.04	<0.01
完整性系数K_v	<0.15	0.15～0.35	0.35～0.55	0.55～0.75	>0.75

下载: 导出CSV

表 8 岩层分布情况

Table 8. Distribution of rock strata

序号	岩土层名称	岩性特征	层厚/m
1	素填土	黄色、紫红色等，结构松散，主要由黏性土夹板岩碎石组成，未完成自重固结	3～4.5
2	粉质黏土	褐黄、褐红色，硬塑状，含少量碎石	0.5～3.5
3	强风化炭质硅质板岩	灰褐、褐黄色，大部分矿物已风化变质，岩质较软，岩体极破碎，岩芯多呈碎片、碎块状，局部风不均匀，残留少量硅质岩块	32～45
4	中等风化炭质硅质板岩	深灰、灰黑色，岩质较硬—坚硬，岩体破碎，岩芯呈碎块状、块状、少量呈短柱状	基岩

下载: 导出CSV

表 9 [BQ]的预测值与实测值对比表

Table 9. Comparison between predicted and measured [BQ] values

编号	隧道桩号	岩石坚硬程度R_c	掌子面完整性系数K_v	地下水K₁	岩体产状K₂	初始地应力K₃	[BQ]值
编号	隧道桩号	岩石坚硬程度R_c	掌子面完整性系数K_v	地下水K₁	岩体产状K₂	初始地应力K₃	预测值	实测值
1	K1+740	20.6	0.3	0.7	0.2	0	146.8	145
2	K1+756	60.6	0.7	0.3	0.2	0	406.8	418
3	K1+764	61.1	0.5	0.3	0.2	0.5	370.8	380
4	K1+770	47.3	0.5	0.7	0.2	0	276.9	270
5	K1+778	60.7	0.48	0.7	0.2	0	312.1	315
6	K1+785	40.5	0.5	0.2	0.2	0	306.5	305
7	K1+791	60.4	0.75	0.3	0.2	0	418.7	418
8	K1+799	21.3	0.3	0.8	0.2	0	138.9	135
9	K1+803	60.2	0.66	0.3	0.2	0.5	345.6	380
10	K1+810	60.1	0.51	0.7	0.2	0	317.8	315

下载: 导出CSV

参考文献(27)

[1]	黄音昊. 隧道围岩分级特征参数的图像智能识别及应用[D]. 成都: 成都理工大学, 2021. (HUANG Y H. Image recognition of Feature parameters of surrounding rock classification in tunnels and its application[D]. Chengdu: Chengdu University of Technology, 2021. (in Chinese) HUANG Y H. Image recognition of Feature parameters of surrounding rock classification in tunnels and its application[D]. Chengdu: Chengdu University of Technology, 2021. (in Chinese)
[2]	TERZAGHI K. Rock defects and loads on tunnel supports[M]//PROCTOR R V, WHITE T L. Rock Tunneling with Steel Supports. Youngstown: Commercial Shearing and Stamping Company, 1946.
[3]	WICKHAM G E, TIEDEMANN H R, SKINNER E H. Support determination based on geologic predictions[C]//North American Rapid Excavation and Tunneling Conference. Chicago: Society of Mining Engineers of the American Institute of Mining, Metallurgical and Petroleum Engineers, 1972: 43-64.
[4]	BIENIAWSKI Z T. Engineering rock mass classifications: a complete manual for engineers and geologists in mining, civil, and petroleum engineering[M]. New York: John Wiley & Sons, 1989.
[5]	BARTON N, LIEN R, LUNDE J. Engineering classification of rock masses for the design of tunnel support[J]. Rock Mechanics, 1974, 6(4): 189-236. doi: 10.1007/BF01239496
[6]	DEERE D U, HENDRON A J, PATTON F D, et al. Design of surface and near-surface construction in rock[C]//The 8th U. S. Symposium on Rock Mechanics (USRMS). Minneapolis, 1966.
[7]	中华人民共和国住房和城乡建设部. 工程岩体分级标准: GB/T 50218—2014[S]. 北京: 中国计划出版社, 2015. (Ministry of Housing and Urban Rural Development of the People’s Republic of China. Standard for engineering classification of rock mass: GB/T 50218—2014[S]. Beijing: China Planning Press, 2015. (in Chinese) Ministry of Housing and Urban Rural Development of the People’s Republic of China. Standard for engineering classification of rock mass: GB/T 50218—2014[S]. Beijing: China Planning Press, 2015. (in Chinese)
[8]	中华人民共和国住房和城乡建设部. 水力发电工程地质勘察规范: GB 50287—2016[S]. 北京: 中国计划出版社, 2017. (Ministry of Housing and Urban Rural Development of the People’s Republic of China. Code for hydropower engineering geological investigation: GB 50287—2016[S]. Beijing: China Planning Press, 2017. (in Chinese) Ministry of Housing and Urban Rural Development of the People’s Republic of China. Code for hydropower engineering geological investigation: GB 50287—2016[S]. Beijing: China Planning Press, 2017. (in Chinese)
[9]	TU W F, LI L P, LI S S, et al. Research on the application of dynamic weighting on the rock mass quality rating[J]. Arabian Journal of Geosciences, 2019, 12(3): 87. doi: 10.1007/s12517-019-4264-9
[10]	姜鹏, 章剑青, 朱征平. 公路隧道围岩分级岩体基本质量指标修正方法探讨[J]. 公路, 2020, 65(1): 332-335. (JIANG P, ZHANG J Q, ZHU Z P. Discussion on the correction method of basic quality indicators for graded rock mass of highway tunnel surrounding rock[J]. Highway, 2020, 65(1): 332-335. (in Chinese) JIANG P, ZHANG J Q, ZHU Z P. Discussion on the correction method of basic quality indicators for graded rock mass of highway tunnel surrounding rock[J]. Highway, 2020, 65（1）: 332-335. (in Chinese)
[11]	童建军, 桂登斌, 王明年, 等. 岩溶隧道围岩级别修正方法研究[J]. 隧道建设(中英文), 2021, 41(S1): 99-107. (TONG J J, GUI D B, WANG M N, et al. Modifying methods for surrounding rock grade of karst tunnel[J]. Tunnel Construction, 2021, 41(S1): 99-107. (in Chinese) TONG J J, GUI D B, WANG M N, et al. Modifying methods for surrounding rock grade of karst tunnel[J]. Tunnel Construction, 2021, 41（S1）: 99-107. (in Chinese)
[12]	吴敏, 黄智, 刘大刚. 基于BQ值修正的岩溶隧道围岩分级方法研究与应用[J]. 高速铁路技术, 2020, 11(6): 1-5. (WU M, HUANG Z, LIU D G. Research and application of surrounding rock classification method of karst tunnel based on BQ value correction[J]. High Speed Railway Technology, 2020, 11(6): 1-5. (in Chinese) WU M, HUANG Z, LIU D G. Research and application of surrounding rock classification method of karst tunnel based on BQ value correction[J]. High Speed Railway Technology, 2020, 11（6）: 1-5. (in Chinese)
[13]	李曙光, 申艳军, 谢江胜, 等. 高原山岭隧道围岩分级方法适宜性评价及初步优化思路[J]. 铁道建筑技术, 2023(8): 33-35,58. (LI S G, SHEN Y J, XIE J S, et al. Suitability evaluation and preliminary optimization of surrounding rock classification method for Plateau Mountain Tunnel[J]. Railway Construction Technology, 2023(8): 33-35,58. (in Chinese) doi: 10.3969/j.issn.1009-4539.2023.08.008 LI S G, SHEN Y J, XIE J S, et al. Suitability evaluation and preliminary optimization of surrounding rock classification method for Plateau Mountain Tunnel[J]. Railway Construction Technology, 2023（8）: 33-35,58. (in Chinese) doi: 10.3969/j.issn.1009-4539.2023.08.008
[14]	邹卫东. 抽水蓄能电站交通隧洞围岩分级方法研究[J]. 铁道建筑技术, 2023(8): 53-58. (ZOU W D. Study on classification method of surrounding rock of traffic tunnel in pumped storage power station[J]. Railway Construction Technology, 2023(8): 53-58. (in Chinese) doi: 10.3969/j.issn.1009-4539.2023.08.013 ZOU W D. Study on classification method of surrounding rock of traffic tunnel in pumped storage power station[J]. Railway Construction Technology, 2023（8）: 53-58. (in Chinese) doi: 10.3969/j.issn.1009-4539.2023.08.013
[15]	樊纯坛, 梁庆国, 岳建平, 等. 层状岩体地下洞室施工阶段围岩精细化分级[J]. 中国公路学报, 2023, 36(4): 169-182. (FAN C T, LIANG Q G, YUE J P, et al. Study on fine classification of surrounding rock during construction stage of underground cavern with layered rock mass[J]. China Journal of Highway and Transport, 2023, 36(4): 169-182. (in Chinese) FAN C T, LIANG Q G, YUE J P, et al. Study on fine classification of surrounding rock during construction stage of underground cavern with layered rock mass[J]. China Journal of Highway and Transport, 2023, 36（4）: 169-182. (in Chinese)
[16]	郑颖人, 阿比尔的. 岩质隧道围岩稳定分析与分级研讨[J]. 现代隧道技术, 2022, 59(1): 1-13. (ZHENG Y R, ABI E D. On Stability analysis and classification of surrounding rocks in rock tunnels[J]. Modern Tunnelling Technology, 2022, 59(1): 1-13. (in Chinese) ZHENG Y R, ABI E D. On Stability analysis and classification of surrounding rocks in rock tunnels[J]. Modern Tunnelling Technology, 2022, 59（1）: 1-13. (in Chinese)
[17]	唐中华, 宋金龙, 潘江波, 等. 基于近景摄影测量的隧道围岩质量快速评价方法研究[J]. 科技通报, 2022, 38(11): 71-77. (TANG Z H, SONG J L, PAN J B, et al. Study on rapid grading of surrounding rock in tunnel based on photogrammetry[J]. Bulletin of Science and Technology, 2022, 38(11): 71-77. (in Chinese) TANG Z H, SONG J L, PAN J B, et al. Study on rapid grading of surrounding rock in tunnel based on photogrammetry[J]. Bulletin of Science and Technology, 2022, 38（11）: 71-77. (in Chinese)
[18]	郭瑞东, 郝付军. 综合超前地质预报在围岩分级的应用研究[J]. 江苏建筑, 2022(3): 118-122. (GUO R D, HAO F J. Application study of comprehensive advance geological prediction in surrounding rock classification[J]. Jiangsu Construction, 2022(3): 118-122. (in Chinese) GUO R D, HAO F J. Application study of comprehensive advance geological prediction in surrounding rock classification[J]. Jiangsu Construction, 2022（3）: 118-122. (in Chinese)
[19]	王元清. 基于超前地质预报技术的隧道围岩分级方法研究[J]. 公路, 2021, 66(3): 398-401. (WANG Y Q. Research on tunnel surrounding rock classification method based on advanced geological prediction technology[J]. Highway, 2021, 66(3): 398-401. (in Chinese) WANG Y Q. Research on tunnel surrounding rock classification method based on advanced geological prediction technology[J]. Highway, 2021, 66（3）: 398-401. (in Chinese)
[20]	鲜晴羽, 仇文革, 王泓颖, 等. 基于卷积神经网络的隧道掌子面图像质量评价方法研究[J]. 铁道科学与工程学报, 2020, 17(3): 563-572. (XIAN Q Y, QIU W G, WANG H Y, et al. Research on image quality assessment method of tunnel face based on convolutional neural network[J]. Journal of Railway Science and Engineering, 2020, 17(3): 563-572. (in Chinese) XIAN Q Y, QIU W G, WANG H Y, et al. Research on image quality assessment method of tunnel face based on convolutional neural network[J]. Journal of Railway Science and Engineering, 2020, 17（3）: 563-572. (in Chinese)
[21]	CHEN J Y, ZHANG D M, HUANG H W, et al. Image-based segmentation and quantification of weak interlayers in rock tunnel face via deep learning[J]. Automation in Construction, 2020, 120: 103371. doi: 10.1016/j.autcon.2020.103371
[22]	ZHANG W J, ZHANG W Q, ZHANG G L, et al. Hard-rock tunnel lithology identification using multi-scale dilated convolutional attention network based on tunnel face images[J]. Frontiers of Structural and Civil Engineering, 2023, 17(12): 1796-1812. doi: 10.1007/s11709-023-0002-1
[23]	彭磊, 周春, 胡锋, 等. 基于深度学习的隧道掌子面节理智能检测与分割[J]. 人民长江, 2023, 54(10): 243-250. (PENG L, ZHOU C, HU F, et al. Intelligent detection and segmentation of tunnel face joints based on deep learning[J]. Yangtze River, 2023, 54(10): 243-250. (in Chinese) PENG L, ZHOU C, HU F, et al. Intelligent detection and segmentation of tunnel face joints based on deep learning[J]. Yangtze River, 2023, 54（10）: 243-250. (in Chinese)
[24]	中华人民共和国交通运输部. 公路隧道设计规范第一册土建工程: JTG 3370.1—2018[S]. 北京: 人民交通出版社, 2019. (Ministry of Transport of the People’s Republic of China. Specifications for design of highway tunnels section 1 civil engineering: JTG 3370.1—2018[S]. Beijing: China Communications Press, 2019. (in Chinese) Ministry of Transport of the People’s Republic of China. Specifications for design of highway tunnels section 1 civil engineering: JTG 3370.1—2018[S]. Beijing: China Communications Press, 2019. (in Chinese)
[25]	王资健. 基于深度学习的隧道内人员防护装备视觉检测技术研究[D]. 长沙: 中南大学, 2022. (WANG Z J. Research on personal protective equipment detection in tunnel construction sites using computer vision and deep learning approaches[D]. Changsha: Central South University, 2022. (in Chinese) WANG Z J. Research on personal protective equipment detection in tunnel construction sites using computer vision and deep learning approaches[D]. Changsha: Central South University, 2022. (in Chinese)
[26]	中华人民共和国住房和城乡建设部. 回弹法检测混凝土抗压强度技术规程: JGJ/T 23—2011[S]. 北京: 中国建筑工业出版社, 2011. (Ministry of Housing and Urban Rural Development of the People’s Republic of China. Technical specification for inspecting of concrete compressive strength by rebound method: JGJ/T 23—2011[S]. Beijing: China Architecture & Building Press, 2011. (in Chinese) Ministry of Housing and Urban Rural Development of the People’s Republic of China. Technical specification for inspecting of concrete compressive strength by rebound method: JGJ/T 23—2011[S]. Beijing: China Architecture & Building Press, 2011. (in Chinese)
[27]	谢飞鸿, 董建辉, 蔡建华, 等. 隧道围岩快速分级及超欠挖爆破智能控制研究[M]. 北京: 科学出版社, 2022. (XIE F H, DONG J H, CAI J H, et al. Research on rapid classification of tunnel surrounding rock and intelligent control of over excavation and under excavation blasting[M]. Beijing: Science Press, 2022. (in Chinese) XIE F H, DONG J H, CAI J H, et al. Research on rapid classification of tunnel surrounding rock and intelligent control of over excavation and under excavation blasting[M]. Beijing: Science Press, 2022. (in Chinese)