摘要 针对隧道掌子面图像质量退化与特征表达单一性问题,提出一种融合多尺度渐进式增强与深度语义建模的围岩智能评估方法。通过多尺度图像增强框架,结合空域滤波与变换域去噪,信噪比提升12.7 dB,显著增强裂隙、节理等关键地质结构的可见性。基于灰度共生矩阵、改进型局部二值模式和RGB颜色矩,构建四维综合评价指标体系,实现岩体状态量化表征。改进ResNet架构集成多尺度特征提取和双重注意力机制(通道注意力+空间注意力),并采用加权交叉熵-标签平滑复合损失函数,解决类别不平衡问题。基于5 000张掌子面图像构建多层数据库,模型在测试集上准确率达94.27%(较基准模型提升4.93%),计算复杂度4.8 G FLOPs(floating point operations)。实际案例验证表明,该方法可为隧道施工提供实时、客观的地质决策支持,显著提升围岩分级智能化水平。
Abstract: To address the issues of image quality degradation and single feature expression in tunnel face images, an intelligent rock mass evaluation method integrating multi-scale progressive enhancement and deep semantic modeling was proposed. A multi-scale image enhancement framework combining spatial domain filtering and transform domain denoising was employed, by which the signal-to-noise ratio was increased by 12.7 dB with significantly enhanced visibility of key geological structures including fractures and joints. A four-dimensional comprehensive evaluation index system was constructed based on the gray-level co-occurrence matrix, improved local binary pattern, and RGB color moments for quantitative rock mass characterization.The ResNet architecture was improved through integration of multi-scale feature extraction and dual attention mechanisms(channel attention + spatial attention), while a weighted cross-entropy-label smoothing composite loss function was adopted to address class imbalance. Based on a constructed database of 5 000 tunnel face images, the model accuracy of 94.27% was achieved on the test set(4.93% higher than the baseline model)with a computational complexity of 4.8 G FLOPs(floating point operations). Practical case verification indicated that the proposed method could provide real-time,objective geological decision support for tunnel construction,significantly improving the intelligence level of rock mass classification.
钟浩, 张永平, 蔡先庆, 袁松, 孔庆轩, 孙浩, 郭胜. 基于多尺度地质特征增强与深度卷积网络的隧道掌子面围岩等级智能评估[J]. 隧道与地下工程灾害防治, 2025, 7(4): 103-114.
ZHONG Hao, ZHANG Yongping, CAI Xianqing, YUAN Song, KONG Qingxuan, SUN Hao, GUO Sheng. Intelligent assessment of surrounding rock grade of tunnel face based on multi-scale geological feature enhancement and deep convolutional network. Hazard Control in Tunnelling and Underground Engineering, 2025, 7(4): 103-114.
[1] 王辉, 刘亚楠, 孟阳. 复杂艰险地区铁路隧道建设安全韧性优化研究[J]. 铁道工程学报, 2024, 41(9): 99-105. WANG Hui, LIU Ya'nan, MENG Yang. Research on safety resilience optimization of railway tunnel construction in complex and challenging areas[J]. Journal of Railway Engineering, 2024, 41(9): 99-105. [2] 蔡晓枘, 喻勇. 对围岩-支护共同作用原理的讨论[J]. 地下空间与工程学报, 2024, 20(增刊1): 54-58. CAI Xiaorui, YU Yong. Discussion on the interaction of surrounding rock and support[J]. Chinese Journal of Underground Space and Engineering, 2024, 20(Suppl.1): 54-58. [3] 何发亮, 谷明成, 王石春. TBM施工隧道围岩分级方法研究[J]. 岩石力学与工程学报, 2002, 21(9): 1350-1354. HE Faliang, GU Mingcheng, WANG Shichun. Study on surrounding rockmass classification of tunnel cut by TBM[J]. Chinese Journal of Rock Mechanics and Engineering, 2002, 21(9): 1350-1354. [4] 张欢, 张世殊, 李天斌, 等. 基于GAPSO-LightGBM的TBM施工隧道围岩等级智能预测方法[J]. 现代隧道技术, 2025, 62(2): 98-109. ZHANG Huan, ZHANG Shishu, LI Tianbin, et al. GAPSO-LightGBM-based intelligent prediction method of surrounding rock grade in TBM tunnelling[J]. Modern Tunnelling Technology, 2025, 62(2): 98-109. [5] 周梦琳, 陈强, 汪波, 等. 基于图像识别的公路隧道围岩智能动态分级研究[J]. 隧道建设(中英文), 2024, 44(6): 1274-1282. ZHOU Menglin, CHEN Qiang, WANG Bo, et al. Intelligent dynamic classification of surrounding rock of highway tunnels based on image recognition[J]. Tunnel Construction(Chinese & English), 2024, 44(6): 1274-1282. [6] XUE Y D, LUO W, CHEN L, et al. An intelligent method for TBM surrounding rock classification based on time series segmentation of rock-machine interaction data[J]. Tunnelling and Underground Space Technology, 2023, 140: 105317. [7] 陈卫东, 李天斌, 黄音昊, 等. 隧道围岩分级特征智能识别及可视化研究[J]. 铁道科学与工程学报, 2024, 21(1): 406-421. CHEN Weidong, LI Tianbin, HUANG Yinhao, et al. Intelligent identification of classification features of tunnel surrounding rock and visualization[J]. Journal of Railway Science and Engineering, 2024, 21(1): 406-421. [8] WANG X Q, GAO H B, JIA Z M. Diffusion-based road defect detection model integrating edge information and efficient multi-scale convolution[J]. Applied Soft Computing, 2025, 178: 113332. [9] 黎江, 许明慧, 张羽. 基于Mask R-CNN的遥感影像土地分割与轮廓提取[J]. 南京信息工程大学学报(自然科学版), 2021, 13(1): 116-123. LI Jiang, XU Minghui, ZHANG Yu. Land segmentation and contour extraction of remote sensing image based on Mask R-CNN[J]. Journal of Nanjing University of Information Science & Technology(Natural Science Edition), 2021, 13(1): 116-123. [10] 吴开兴, 宋剑. 基于灰度共生矩阵的煤与矸石自动识别研究[J]. 煤炭工程, 2016, 48(2): 98-101. WU Kaixing, SONG Jian. Automatic coal-gangue identification based on gray level co-occurrence matrix[J]. Coal Engineering, 2016, 48(2): 98-101. [11] 熊丽婷, 沈克永, 熊书兴. 鲁棒局部二值模式及其在图像检索中的应用[J]. 计算机工程与设计, 2017, 38(8): 2184-2189. XIONG Liting, SHEN Keyong, XIONG Shuxing. Robust local binary pattern and its application in image retrieval system[J]. Computer Engineering and Design, 2017, 38(8): 2184-2189. [12] IMANI M. Attention based multi-level and multi-scale convolutional network for PolSAR image classification[J]. Advances in Space Research, 2025, 75(11): 7971-7986. [13] 林杰涛, 胡克勇, 袁天龙, 等. 高效注意力与多尺度特征聚合的遥感影像变化检测[J]. 北京邮电大学学报, 2025, 48(4): 59-64. LIN Jietao, HU Keyong, YUAN Tianlong, et al. Sensing Image Change Detection with Efficient Attention and Multi-Scale Feature Aggregation[J]. Journal of Beijing University of Posts and Telecommunications, 2025, 48(4): 59-64. [14] SINHA M, PAUL S, NEE LALA M G. Comparative analysis of mask R-CNN and U-Net architectures using ResNet as backbone for lunar crater detection[J]. Planetary and Space Science, 2025, 264: 106140. [15] ZHAO X, TANG C, HU H S, et al. Attention mechanism based multimodal feature fusion network for human action recognition[J]. Journal of Visual Commu-nication and Image Representation, 2025, 110: 104459. [16] 史崇崴, 张群, 刘治东, 等. 基于灰度共生矩阵的微动干扰分类识别[J]. 空军工程大学学报, 2022, 23(4): 35-42. SHI Chongwei, ZHANG Qun, LIU Zhidong, et al. A micro-motion jamming classification and recognition method based on gray-level co-occurrence matrix[J]. Journal of Air Force Engineering University, 2022, 23(4): 35-42. [17] 贾小云, 潘德燃. 基于最大稳定极值区域与主要旋转不变局部二值模式的复杂场景车牌检测方法[J]. 科学技术与工程, 2020, 20(4): 1509-1513. JIA Xiaoyun, PAN Deran. License plate detection in complex scene based on maximum stable extremum region and dominant rotation invariant local binary mode[J]. Science Technology and Engineering, 2020, 20(4): 1509-1513. [18] 张少博, 全书海, 石英, 等. 基于颜色矩的图像检索算法研究[J]. 计算机工程, 2014, 40(6): 252-255. ZHANG Shaobo, QUAN Shuhai, SHI Ying, et al. Study on image retrieval algorithm based on color moment[J]. Computer Engineering, 2014, 40(6): 252-255. [19] 熊丽婷, 沈克永, 熊书兴. 鲁棒局部二值模式及其在图像检索中的应用[J]. 计算机工程与设计, 2017, 38(8): 2184-2189. XIONG Liting, SHEN Keyong, XIONG Shuxing. Robust local binary pattern and its application in image retrieval system[J]. Computer Engineering and Design, 2017, 38(8): 2184-2189. [20] 宋琳琳, 王相海. 一种综合颜色和纹理特征的图像检索算法[J]. 计算机工程与应用, 2011, 47(34): 203-207. SONG Linlin, WANG Xianghai. Image retrieval algorithm using both color and texture features[J]. Computer Engineering and Applications, 2011, 47(34): 203-207. [21] 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: 12-16.
2025, Vol. 7 
