Rock mass structural characterization based on machine vision and grout take prediction
OUYANG Shaoming1, DING Changdong1, 2*, ZHOU Liming1, DING Xiang3, LUO Rong1, HU Dawei2, 4
(1. Key Laboratory of Geotechnical Mechanics and Engineering of the Ministry of Water Resources, Changjiang River Scientific Research Institute, Wuhan, Hubei 430010, China; 2. State Key Laboratory of Geomechanics and Geotechnical Engineering Safety, Institute of Rock and Soil Mechanics, Chinese Academy of Sciences, Wuhan, Hubei 430071, China; 3. School of Civil Engineering and Architecture, Hubei University of Technology, Wuhan, Hubei 430068, China; 4. State Key Laboratory of Deep Geothermal Resources, China University of Geosciences, Wuhan, Hubei 430074, China)
Abstract：Aiming at “black-box” technical bottleneck characterized by the inherent concealment and evaluation challenges in underground cavern grouting, this study proposes an integrated framework combining deep learning and machine learning for refined characterization of rock mass structures and quantitative assessment of grouting performance. The results indicated: (1) By employing high-resolution image acquisition and 3D point-cloud reconstruction technologies to construct a dataset of the surrounding rock, an efficient fracture identification and segmentation method based on the CBAM-UKAN model was developed. The proposed method achieved IoU and Dice coefficients of 0.76 and 0.86, respectively, on the validation set. (2) Through the integration of morphological closing, skeleton extraction, and a multi-scale geometric quantification algorithm chain, key fracture parameters such as length, width, and area were extracted, providing essential inputs for the rock mass structural characterization. (3) Comparative analysis of four ensemble learning models identified the simulated annealing-optimized Gradient Boosting Regression Tree (SA-GBRT) as the best-performing predictive model, achieving an R² value of 0.79 on the test set. The proposed approach enables reliable prediction of unit ash consumption based on rock mass fracture characteristics and grouting construction parameters, thereby providing a quantitative basis for pre-evaluation of seepage-control performance. These findings offer significant practical value for the intelligent optimization and scientific decision-making of grouting operations in underground engineering.
欧阳劭明1，丁长栋1，2*，周黎明1，丁 祥3，罗 荣1，胡大伟2，4. 基于机器视觉的岩体结构表征与灌浆量预测研究[J]. 岩石力学与工程学报, 2026, 45(S1): 282-296.
OUYANG Shaoming1, DING Changdong1, 2*, ZHOU Liming1, DING Xiang3, LUO Rong1, HU Dawei2, 4. Rock mass structural characterization based on machine vision and grout take prediction. , 2026, 45(S1): 282-296.
[1] 谢和平，高明忠，张 茹，等. 地下生态城市与深地生态圈战略构想及其关键技术展望[J]. 岩石力学与工程学报，2017，36(6)：1 301–1 313. (XIE Heping，GAO Mingzhong，ZHANG Ru，et al. The subversive idea and its key technical prospect on underground ecological city and ecosystem[J]. Chinese Journal of Rock Mechanics and Engineering，2017，36(6)：1 301–1 313.(in Chinese))
[2] HULME W T，ZHAO J. Underground space development in Singapore-the past，present and future[J]. Tunnelling and Underground Space Technology，1999，14(4)：407.
[3] 李术才，李树忱，张庆松，等. 岩溶裂隙水与不良地质情况超前预报研究[J]. 岩石力学与工程学报，2007，26(2)：217–225.(LI Shucai，LI Shuchen，ZHANG Qingsong，et al. Forecast of karst-fractured groundwater and defective geological conditions[J]. Chinese Journal of Rock Mechanics and Engineering，2007，26(2)：217–225.(in Chinese))
[4] 蒲春平，孙耀南. 隧道与地下工程渗漏水现状及其防治措施综述[J]. 世界隧道，1999，(3)：23–27.(PU Chunping，SUN Yaonan. Water leakage problem and its countermeasures for tunneling and underground works[J]. World Tunneling，1999，(3)：23–27.(in Chinese))
[5] XIAO F，LIU Q，ZHAO Z Y，et al. Information and knowledge behind data from underground rock grouting[J]. Journal of Rock Mechanics and Geotechnical Engineering，2021，13(6)：1 326–1 339.
[6] BARTON N. Rock quality，seismic velocity，attenuation and anisotropy[M]. 1st ed. [S. l.]：CRC Press，2006：9–12.
[7] 喻春平，郭 强，廖正彪，等. 岩石表面图像的特征提取和处理方法[J]. 上海交通大学学报，2013，47(9)：1 482–1 486.(YU Chunping，GUO Qiang，LIAO Zhengbiao，et al. Feature extraction and corresponding processing of rock surface image[J]. Journal of Shanghai Jiaotong University，2013，47(9)：1 482–1 486.(in Chinese))
[8] 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.(CHEN Xiangsheng，FU Yanbin，CHEN Xi，et al. Progress in underground space construction technology and technical challenges of digital intelligence[J]. China Journal of Highway and Transport，2022，35(1)：1–12.(in Chinese))
[9] 朱合华，潘柄屹，武 威，等. 岩体结构面信息采集及识别方法研究进展[J]. 应用基础与工程科学学报，2023，31(6)：1 339–1 360.(ZHU Hehua，PAN Bingyi，WU Wei，et al. Review on collection and extraction methods of rock mass discontinuity information[J]. Journal of Basic Science and Engineering，2023，31(6)：1 339–1 360.(in Chinese))
[10] YUAN Y，ZHANG N，HAN C，et al. Automated identification of fissure trace in mining roadway via deep learning[J]. Journal of Rock Mechanics and Geotechnical Engineering，2023，15(8)：2 039–2 052.
[11] LI M Z，CHEN M，LU W B，et al. Automatic extraction and quantitative analysis of characteristics from complex fractures on rock surfaces via deep learning[J]. International Journal of Rock Mechanics and Mining Sciences，2025，187：106038.
[12] CHEN J，ZHOU M，HUANG H，et al. Automated extraction and evaluation of fracture trace maps from rock tunnel face images via deep learning[J]. International Journal of Rock Mechanics and Mining Sciences，2021，142：104745.
[13] 吴 金，吴顺川，王 焘，等. 基于深度学习的复杂地层钻孔图像岩体结构面识别方法[J]. 清华大学学报：自然科学版，2024，64(7)：1 136–1 146.(WU Jin，WU Shunchuan，WANG Tao，et al. Deep learning-based method for rock discontinuity recognition in complex stratum borehole images[J]. Journal of Tsinghua University：Science and Technology，2024，64(7)：1 136–1 146.(in Chinese))
[14] 李元海，徐晓华，朱鸿鹄，等. 基于计算机视觉的岩石裂隙识别表征与软件研制[J]. 岩土工程学报，2024，46(3)：459–469.(LI Yuanhai，XU Xiaohua，ZHU Honghu，et al. Identification and characterization of rock fractures based on computer vision and software development[J]. Chinese Journal of Geotechnical Engineering，2024，46(3)：459–469.(in Chinese))
[15] MEHRISHAL S，KIM J，SHAO Y，et al. Artificial intelligence-aided semi–automatic joint trace detection from textured three-dimensional models of rock mass[J]. Journal of Rock Mechanics and Geotechnical Engineering，2025，17(4)：1 973–1 985.
[16] 薛东杰，唐麒淳，王 傲，等. 基于FCN的岩石混凝土裂隙几何智能识别[J]. 岩石力学与工程学报，2019，38(增2)：3 393–3 403.(XUE Dongjie，TANG Qichun，WANG Ao，et al. FCN-based intelligent identification of crack geometry in rock or concrete[J]. Chinese Journal of Rock Mechanics and Engineering，2019，38(Supp.2)：3 393– 3 403.(in Chinese))
[17] LOMBARDI G，DEERE D. Grouting design and control using the GIN principle[J]. Water Power and Dam Construction，1993，45(6)：15–22.
[18] 张宜虎，张立德，赵 晓，等. 水封洞库防渗注浆工程量估算方法研究[J]. 岩石力学与工程学报，2025，44(增1)：42–50.(ZHANG Yihu，ZHANG Lide，ZHAO Xiao，et al. Research on the estimation method of anti-seepage grouting engineering quantity for water-sealed caverns[J]. Chinese Journal of Rock Mechanics and Engineering，2025，44(Supp.1)：42–50.(in Chinese))
[19] DU J J，LIU J Q，LYU X Z，et al. Analytical solution of the grouting reinforcement response of a circular cavern in deeply buried fractured surrounding rock under a nonuniform stress field[J]. ACS Omega，2022，7(32)：28 016–28 029.
[20] DING W Q，LEI B，DUAN C，et al. Numerical simulation of grout diffusion in rough rock fractures considering multiple influencing factors[J]. Rock Mechanics and Rock Engineering，2025，58(1)：355–377.
[21] ERIKSSON M，STILLE H，ANDERSSON J，et al. Numerical calculations for prediction of grout spread with account for filtration and varying aperture[J]. Tunnelling and Underground Space Technology，2000，15(4)：353–364.
[22] YANG C P. Estimating cement take and grout efficiency on foundation improvement for Li-Yu-Tan dam[J]. Engineering Geology，2004，75(1)：1–14.
[23] 祝玉珊，王晓玲，崔 博，等. 坝基灌浆量预测ISSA-Stacking集成学习代理模型研究[J]. 天津大学学报：自然科学与工程技术版，2024，57(2)：174–185.(ZHU Yushan，WANG Xiaoling，CUI Bo，et al. Research on an ISSA-Stacking ensemble learning surrogate model of dam foundation grouting volume prediction[J]. Journal of Tianjin University：Science and Technology，2024，57(2)：174–185.(in Chinese))
[24] 李 凯，任炳昱，关 涛，等. 帷幕灌浆量区间预测的Bootstrap-IGWO-SVM模型研究[J]. 水力发电学报，2022，41(10)：18–29. (LI Kai，REN Bingyu，GUAN Tao，et al. Curtain grouting cement interval prediction using Bootstrap-IGWO-SVM model[J]. Journal of Hydroelectric Engineering，2022，41(10)：18–29.(in Chinese))
[25] RAHIMI E，TESHNIZI E S，RASTEGARNIA A，et al. Cement take estimation using neural networks and statistical analysis in Bakhtiari and Karun 4 dam sites，in south west of Iran[J]. Bulletin of Engineering Geology and the Environment，2019，78(4)：2 817–2 834.
[26] WOO S，PARK J，LEE J Y，et al. CBAM：convolutional block attention module[C]// Proceedings of the European Conference on Computer Vision(ECCV). Springer，Cham：[s. n.]，2018：3–19.
[27] LIU Z，WANG Y，VAIDYA S，et al. KAN：Kolmogorov-Arnold networks[C]// The 13th International Conference on Learning Representations. [S. l.]：[s. n.]，2025：66 342–66 388.
[28] 薛家祥，刘 晓，朱思君，等. 基于分形和数学形态学的TIG焊熔池图像分析[J]. 华南理工大学学报：自然科学版，2007，35(8)：7–10.(XUE Jiaxiang，LIU Xiao，ZHU Sijun，et al. Analysis of molten pool image of TIG welding based on fractal and mathematical morphology[J]. Journal of South China University of Technology：Natural Science，2007，35(8)：7–10.(in Chinese))
[29] ZHANG T Y，SUEN C Y. A fast parallel algorithm for thinning digital patterns[J]. Communications of the ACM，1984，27(3)：236–239.
[30] LORENZI L，MELGANI F，MERCIER G. Inpainting strategies for reconstruction of missing data in VHR images[J]. IEEE Geoscience and Remote Sensing Letters，2011，8(5)：914–918.
[31] DENG S H，WANG X L，ZHU Y S，et al. Hybrid grey wolf optimization algorithm-based support vector machine for groutability prediction of fractured rock mass[J]. Journal of Computing in Civil Engineering，2019，33(2)：04018065.
[32] DING C D，OUYANG S M，ZHANG L D，et al. Application of machine learning techniques for predicting grouting performance in underground caverns considering rock structure[J]. Engineering Geology，2026，360：108485.
[33] 丁长栋，张宜虎，李 领，等. 地下水封洞库帷幕灌浆试验与渗控效果分析[J]. 隧道与地下工程灾害防治，2024，6(1)：64–72.(DING Changdong，ZHANG Yihu，LI Ling，et al. Curtain grouting test and seepage control effect analysis of underground water-sealed oil storage[J]. Hazard Control in Tunnelling and Underground Engineering，2024，6(1)：64–72.(in Chinese))
[34] 欧阳劭明，丁长栋，丁 祥，等. 基于TPE–GBT模型的地下水封洞库灌浆量预测研究[J]. 现代隧道技术，2024，61(5)：138–145.(OUYANG Shaoming，DING Changdong，DING Xiang，et al. Research on grouting volume prediction for underground water-sealed caverns based on TPE-GBT model[J]. Modern Tunnelling Technology，2024，61(5)：138–145.(in Chinese))
[35] PEDREGOSA F，VAROQUAUX G，GRAMFORT A，et al. Scikit-learn：machine learning in python[J]. Journal of Machine Learning Research，2011，12：2 825–2 830.
[36] 刘华强，陆明志，殷宗泽，等. 基于模拟退火算法的边坡临界滑面搜索方法[J]. 岩石力学与工程学报，2008，27(增2)：3 686–3 691.(LIU Huaqiang，LU Mingzhi，YIN Zongze，et al. Analytical method of searching noncircular slip surface based on simulated annealing optimization algorithm[J]. Chinese Journal of Rock Mechanics and Engineering，2008，27(Supp.2)：3 686–3 691.(in Chinese))
2026, Vol. 45 
