Image-based dominant fracture recognition for cloud-model blastability classification in underground stopes: Development and field validation

Abstract

The difficulty in accurately quantifying fracture features at underground excavation faces constrains stope blastability evaluation and refined blasting operations. An intelligent workflow—fracture recognition, indicator quantification, blastability classification, and scheme matching—was developed by integrating a You Only Look Once (YOLO)-based segmentation model with a comprehensive cloud model. Multimodal data augmentation was designed to emulate harsh underground imaging conditions, expanding 119 labeled images to 474 for model training. With DeepLabV3 and YOLOv8l-seg as reference models, YOLOv5-seg variants were benchmarked using Dice, pixel accuracy (PA), mask intersection over union (Mask IoU), and box intersection over union (Box IoU). YOLOv5x-seg was then selected and further enhanced by replacing PANet with a bidirectional feature pyramid network (BiFPN) and embedding a convolutional block attention module (CBAM), yielding YOLOv5x-CBF with consistent metric gains. Fracture trace length was quantified within the effective face region and integrated with powder factor and blasting advance. These indicators were then used to construct a comprehensive cloud model for uncertainty representation, three-class blastability classification, and “class–scheme” mapping. Field-scale trials on two stopes demonstrated improved blasting performance, with blasting advance increased by 18%–30% and powder factor reduced by 8%–13%. These results confirm the practicality of the proposed method and provide a field-deployable technical reference for intelligent blasting in underground metal mines.

Keywords

underground metal mine / blastability classification / intelligent mining / face fracture segmentation

References

[1] Wang YM, Chen QS, Dai BB, Wang DL. Guidance and review: Advancing mining technology for enhanced production and supply of strategic minerals in China. Green Smart Min. Eng., 2024, 112

[2] Li P, Cai MF, Miao SJ, Ren FH, Gorjian M, Peng C. Mechanism, prevention, and control of mining-induced dynamic disasters in underground metal mines in China: Challenges and solutions. J. Cent. South Univ., 2024, 3182549

[3] Q. Na, Q.S. Chen, D.L. Wang, J.L. Pan, Y.B. Tao, and A.X. Wu, A novel non-destructive testing method for the uniaxial compressive strength of cemented paste backfill based on hyperspectral imaging and artificial intelligence, NDT E Int., 160(2026), art. No. 103654.

[4] Li XB, Gong FQ, Tao M, et al. . Failure mechanism and coupled static-dynamic loading theory in deep hard rock mining: A review. J. Rock Mech. Geotech. Eng., 2017, 94767

[5] Xu P, Yang RS, Zuo JJ, et al. . Research progress of the fundamental theory and technology of rock blasting. Int. J. Miner. Metall. Mater., 2022, 294705

[6] Z.X. Hong, M. Tao, C.Q. Wu, J. Zhou, and D. Wang, The spatial distribution of excavation damaged zone around underground roadways during blasting excavation, Bull. Eng. Geol. Environ., 82(2023), No. 4, art. No. 155.

[7] Zhang ZX, Sanchidrián JA, Ouchterlony F, Luukkanen S. Reduction of fragment size from mining to mineral processing: A review. Rock Mech. Rock Eng., 2023, 561747

[8] E.M. Kinyua, J.H. Zhang, R.M. Kasomo, D. Mauti, and J. Mwangangi, A review of the influence of blast fragmentation on downstream processing of metal ores, Miner. Eng., 186(2022), art. No. 107743.

[9] Norgate T, Haque N. Energy and greenhouse gas impacts of mining and mineral processing operations. J. Clean. Prod., 2010, 183266

[10] Z.L. Zhou, C.G. Chen, X. Cai, and P.Y. Wang, Blastability evaluation for rock mass: Review and new tendency, Bull. Eng. Geol. Environ., 83(2024), No. 12, art. No. 517.

[11] Jackson J, Gaunt J, Astorga MDimitrakopoulos R. Predicting mill ore feed variability using integrated geotechnical/geometallurgical models. Advances in Applied Strategic Mine Planning.. 2018, Cham, Springer465

[12] X. Liu, P. Yan, W.B. Lu, et al., Numerical investigation of an improved deep-hole presplitting method based on notched blasting for deep-buried high sidewall structures, J. Build. Eng., 70(2023), art. No. 106310.

[13] Wu HJ, Gong M, Yang RS, Wu XD, Liu XY. Double-face intelligent hole position planning method for precision blasting in roadways using a computer-controlled drill jumbo. Int. J. Miner. Metall. Mater., 2023, 3061025

[14] P.K. Luo, X.G. Fang, D.Y. Li, et al., Evaluation of excavation method on point load strength of rocks with poor geological conditions in a deep metal mine, Geomech. Geophys. Geo-Energy Geo-Resour., 9(2023), No. 1, art. No. 90.

[15] Bieniawski Z T. Engineering classification of jointed rock masses. Civ. Eng. Sivil. Ing., 1973, 12335

[16] Barton N, Lien R, Lunde J. Engineering classification of rock masses for the design of tunnel support. Rock Mech., 1974, 64189

[17] Hoek E, Brown E T. Underground Excavations in Rock. 1980, London, CRC Press

[18] Goel RK, Jethwa JL, Paithankar AG. Indian experiences with Q and RMR systems. Tunn. Undergr. Space Technol., 1995, 10197

[19] E.F. Salmi and E.J. Sellers, A review of the methods to incorporate the geological and geotechnical characteristics of rock masses in blastability assessments for selective blast design, Eng. Geol., 281(2021), art. No. 105970.

[20] Li JX, Wu W, Pan BY, Dong BS, Zhu HH. Impact of joint distribution within hard rock mass on tunnel blasting profile: Numerical simulation and field measurement. J. Rock Mech. Geotech. Eng., 2026, 1853760

[21] Wang AA, Cao S, Yilmaz E, Cao H. Research on the visualization method of lithology intelligent recognition based on deep learning using mine tunnel images. Int. J. Miner. Metall. Mater., 2026, 331141

[22] Y.B. Tao, Q.L. Zhang, Q.S. Chen, C.C. Qi, and Y.K. Liu, An intelligent approach using micro-seismic monitoring signal clustering and an optimized K-means model to guide the selection of support patterns in underground mines, Tunn. Undergr. Space Technol., 154(2024), art. No. 106095.

[23] Z.D. Liu, Z.X. Zhang, Q.Q. Zhu, P. Xiao, J.M. Yang, and D.Y. Li, Energy analysis and microfracture characteristics of granite under dynamic modes I and II loading based on image processing techniques, Int. J. Rock Mech. Min. Sci., 195(2025), art. No. 106287.

[24] Xu SY, Wang J, Shou WC, Ngo T, Sadick AM, Wang XY. Computer vision techniques in construction: A critical review. Arch. Comput. Methods Eng., 2021, 2853383

[25] Q. Yuan, Y.F. Shi, and M.Y. Li, A review of computer vision-based crack detection methods in civil infrastructure: Progress and challenges, Remote. Sens., 16(2024), No. 16, art. No. 2910.

[26] Q.L. Zhang, H.S. Wang, J.L. Pan, Y.B. Tao, and Y. Feng, Enhancing SCSegamba with fracture orientation consistency loss for robust rock fracture segmentation in deep underground mining, Appl. Sci., 16(2026), No. 10, art. No. 4717.

[27] Y. Zhang, J.Q. Chen, Y.L. Li, and B. Li, Automatic estimation of RQD based on deep ensemble learning and fracture fitting, Geoenergy Sci. Eng., 230(2023), art. No. 212132.

[28] F. Guo, Y. Qian, J. Liu, and H.Y. Yu, Pavement crack detection based on transformer network, Autom. Constr., 145(2023), art. No. 104646.

[29] Letshwiti TM, Shahsavar M, Moniri-Morad A, Sattarvand J. Deep learning-based image segmentation for highwall stability monitoring in open pit mines. J. Eng. Res., 2025, 1343595

[30] Cai MF, Li P, Tan WH, Ren FH. Key engineering technologies to achieve green, intelligent, and sustainable development of deep metal mines in China. Engineering, 2021, 7111513

[31] C.D. Ma, K.C. Li, J.L. Pan, J.S. Zheng, Q.L. Zhang, and C.C. Qi, The segmentation of tunnel faces in underground mines based on the optimized YOLOv5, Minerals, 15(2025), No. 3, art. No. 255.

[32] Gheorghe C, Duguleana M, Boboc RG, Postelnicu CC. Analyzing real-time object detection with YOLO algorithm in automotive applications: A review. CMES-Comput. Model. Eng. Sci., 2024, 14131939

[33] W.J. Huang, R. Pang, Y. Zhou, F.Y. Zhong, J.S. Chen, and F. Kang, Multi-condition dam crack segmentation based on a novel lightweight CEE-YOLO model, Adv. Eng. Inform., 68(2025), art. No. 103697.

[34] Y.Y. Zhang, J.C. Liu, Y.S. Zhao, J.H. Xu, and W.Z. Wu, Crack detection on concrete composite slab using you only look once version 8 based lightweight system, J. Build. Eng., 108(2025), art. No. 112974.

[35] G. Xu, Q.R. Yue, and X.G. Liu, Deep learning algorithm for real-time automatic crack detection, segmentation, qualification, Eng. Appl. Artif. Intell., 126(2023), art. No. 107085.

[36] R. Khanam and M. Hussain, What is YOLOv5: A deep look into the internal features of the popular object detector, arXiv, 2024, [2025-06-03]. https://arxiv.org/abs/2407.20892.

[37] Z. Qu, L.Y. Gao, S.Y. Wang, H.N. Yin, and T.M. Yi, An improved YOLOv5 method for large objects detection with multiscale feature cross-layer fusion network, Image Vis. Comput., 125(2022), art. No. 104518.

[38] Peng XB, Yin ZJ, Yang Z. Deeplab_v3_plus-net for image semantic segmentation with channel compression. International Conference on Communication Technology Proceedings, ICCT. 20201320

[39] Y.Y. Liu, X.T. Bai, J.F. Wang, G.N. Li, J. Li, and Z.M. Lv, Image semantic segmentation approach based on DeepLabV3 plus network with an attention mechanism, Eng. Appl. Artif. Intell., 127(2024), art. No. 107260.

[40] Q.Y. Wang, L.Y. Guo, and B. Chen, The lightweight CS-YOLO model applied for concrete crack segmentation, Expert Syst. Appl., 277(2025), art. No. 127298.

[41] J. Doherty, B. Gardiner, E. Kerr, and N. Siddique, BiFPNYOLO: One-stage object detection integrating bi-directional feature pyramid networks, Pattern Recognit., 160(2025), art. No. 111209.

[42] Chen J, Mai HS, Luo LB, Chen XQ, Wu KL. Effective feature fusion network in BIFPN for small object detection. Proceedings – International Conference on Image Processing, ICIP. 2021699

[43] Woo S, Park J, Lee JY, Kweon IS, et al. Ferrari V, Hebert M, Sminchisescu C, et al. . CBAM: Convolutional block attention module. Computer Vision–ECCV 2018.. 2018, Cham, Springer3

[44] Zhao ZX, Chen KX, Yamane S. CBAM-Unet++: Easier to find the target with the attention module “cBAM”. 2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021. 2021655

[45] Dice LR. Measures of the amount of ecologic association between species. Ecology, 1945, 263297

[46] Guindon B, Zhang Y. Application of the dice coefficient to accuracy assessment of object-based image classification. Can. J. Remote. Sens., 2017, 43148

[47] D.R. Li, Z.D. Duan, X.Y. Hu, and D.C. Zhang, Pixel-level recognition of pavement distresses based on U-net, Adv. Mater. Sci. Eng., 2021(2021), No. 1, art. No. 5586615.

[48] Cheng BW, Girshick R, Dollar P, Berg AC, Kirillov A. Boundary IoU: Improving object-centric image segmentation evaluation. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. 202115329

[49] L. Yang, K. Zhang, J.Q. Liu, and C.K. Bi, Location IoU: A new evaluation and loss for bounding box regression in object detection, [in] Proceedings of the International Joint Conference on Neural Networks. Yokohama, 2024, art. No. 1.

[50] Li D, Cheung D, Shi XM, Ng V. Uncertainty reasoning based on cloud models in controllers. Comput. Math. Appl., 1998, 35399

[51] G.Q. Huang and L.M. Xiao, Failure mode and effect analysis: An interval-valued intuitionistic fuzzy cloud theory-based method, Appl. Soft Comput., 98(2021), art. No. 106834.

[52] T. Wang, X.T. Wang, L. Wang, C.P. Au-Yong, and A.S. Ali, Assessment of the development level of regional industrialized building based on cloud model: A case study in Guangzhou, China, J. Build. Eng., 44(2021), art. No. 102547.

[53] D.S. Guo, X.X. Yang, J.C. Yang, H.N. Wu, F.Y. Meng, and R.P. Chen, Assessing subway station flood risks with a resilience Framework: Combining weighting methods and two-dimensional cloud model, Tunnelling Underground Space Technol., 168(2026), art. No. 107170.

[54] Y. Liu, H.Y. Chen, and X.J. Wang, Research on green renovations of existing public buildings based on a cloud model–TOPSIS method, J. Build. Eng., 34(2021), art. No. 101930.

[55] H.B. Liang, X.G. Xie, X.X. Chen, et al., Study on risk assessment of tunnel construction across mined-out region based on combined weight-two-dimensional cloud model, Sci. Rep., 15(2025), art. No. 7233.

[56] L. Li, B. Ni, Y. Qiang, S.X. Zhang, D.S. Zhao, and L. Zhou, Risk assessment of debris flow disaster based on the cloud model: Probability fusion method, PLoS One, 18(2023), No. 2, art. No. e0281039.

RIGHTS & PERMISSIONS

University of Science and Technology Beijing