2025年04月05日 星期六

广西师范大学学报（自然科学版） ›› 2025, Vol. 43 ›› Issue (2): 107-120.doi: 10.16088/j.issn.1001-6600.2024010902

• 智能信息处理 • 上一篇    下一篇

基于迁移学习及通道先验注意力机制的地质构造识别

刘俊杰1,2, 马凯1,2, 黄泽华1,2, 田苗3, 邱芹军4,5*, 陶留锋4,5, 谢忠4,5   

  1. 1.湖北省水电工程智能视觉监测重点实验室(三峡大学), 湖北 宜昌 443002;
    2.三峡大学 计算机与信息学院, 湖北 宜昌 443002;
    3.地质探测与评估教育部重点实验室(中国地质大学(武汉)), 湖北 武汉 430074;
    4.中国地质大学(武汉) 计算机学院, 湖北 武汉 430074;
    5.地理信息系统国家地方联合工程实验室(中国地质大学(武汉)), 湖北 武汉 430074
  • 收稿日期:2024-01-09 修回日期:2024-03-27 出版日期:2025-03-05 发布日期:2025-04-02
  • 通讯作者: 邱芹军(1988—), 男, 湖北武汉人, 中国地质大学(武汉)副研究员, 博士。E-mail: qiuqinjun@cug.edu.cn
  • 基金资助:
    国家自然科学基金(42301492); 国家重点研发计划项目(2022YFB3904200, 2022YFF0711601); 湖北省自然科学基金(2022CFB640); 地质探测与评估教育部重点实验室主任基金(GLAB2023ZR01)

Geological Structure Recognition Based on Transfer Learning and Channel Prior Attention Mechanism

LIU Junjie1,2, MA Kai1,2, HUANG Zehua1,2, TIAN Miao3, QIU Qinjun4,5* , TAO Liufeng4,5, XIE Zhong4,5   

  1. 1. Hubei Key Laboratory of Intelligent Visual Monitoring for Hydropower Engineering (China Three Gorges University), Yichang Hubei 443002, China;
    2. College of Computer and Information Technology, China Three Gorges University, Yichang Hubei 443002, China;
    3. Key Laboratory of Geological Survey and Evaluation of Ministry of Education (China University of Geosciences), Wuhan Hubei 430074, China;
    4. School of Computer Science, China University of Geosciences, Wuhan Hubei 430074, China;
    5. National and Local Joint Engineering Laboratory of Geographic Information Systems (China University of Geosciences), Wuhan Hubei 430074, China
  • Received:2024-01-09 Revised:2024-03-27 Online:2025-03-05 Published:2025-04-02

PDF (PC)

0

摘要： 针对平面地质图件中地质构造背景复杂、符号表示多样化而导致识别效果不佳的问题,本文提出一种基于迁移学习和通道先验注意力机制的地质构造识别模型MsAttenEfficientNet。该模型以EfficientNet为主干网络架构,并使用通道先验注意力(channel prior convolution attention, CPCA)模块替换EfficientNet特征提取模块MBConv中的压缩和激励网络(squeeze-and-excitation net, SENet),使模型能够动态地分配通道和空间注意力权重,更准确地捕捉到图像中的重要区域和空间结构;其次对顶层预测模块进行改进,引入Swish激活函数和Dropout层,加强模型的泛化性能;最后使用Adam优化算法提高网络的收敛速度,并利用迁移学习实现特征参数共享。通过在地质构造数据集GeoStr18上进行训练及测试,实验结果表明,MsAttenEfficientNet模型对地质构造的识别精准率为96.92%,召回率为96.89%,F1分数为96.90%,优于ResNet50、ShuffleNetV2和DenseNet121等主流分类识别模型,可有效用于地质构造识别。

关键词: 图像识别, 地质构造, EfficientNet网络, 通道先验注意力, 迁移学习

Abstract: To address the issue of poor identification resulting from the complex geological background and diverse symbol representations in plane geological drawings, a geological structure recognition model, named MsAttenEfficientNet, is proposed based on transfer learning and the Channel Prior Convolution Attention (CPCA) mechanism. In this model, EfficientNet is utilized as the backbone network architecture, and the Squeeze and Excitation Net (SENet) in the MBConv feature extraction module of EfficientNet are replaced by the CPCA module. This replacement enables the dynamic allocation of channel and spatial attention weights, allowing the model to more accurately capture important regions and spatial structures in the images. Subsequently, improvements are made to the top prediction module by introducing the Swish activation function and Dropout layers to enhance the model’s generalization performance. Finally, the Adam optimization algorithm is employed to improve the convergence speed of the network, and transfer learning is utilized to achieve feature parameter sharing. Experimental results, obtained from training and testing on the geological structure dataset GeoStr18, demonstrate that the MsAttenEfficientNet model achieves a precision of 96.92%, a recall of 96.89%, and an F1 score of 96.90% in geological structure recognition, outperforming mainstream classification models such as ResNet50, ShuffleNetV2, and DenseNet121, thus effectively applicable for geological structure recognition.

Key words: image recognition, geological structure, EfficientNet network, CPCA, transfer learning

中图分类号:  TP183; TP391.41

[1] GENG Z, WANG Y F. Physics-guided deep learning for predicting geological drilling risk of wellbore instability using seismic attributes data[J]. Engineering Geology, 2020, 279: 105857. DOI: 10.1016/j.enggeo.2020.105857.
[2] BULKAN S, VANNUCCHI P, GASPERINI L, et al. Modelling tectonic deformation along the North-Anatolian Fault in the Sea of Marmara[J]. Tectonophysics, 2020, 794: 228612. DOI: 10.1016/j.tecto.2020.228612.
[3] 王淑鹏. 野外地质露头图像构造自动识别方法研究[D]. 北京: 中国地质大学(北京), 2021. DOI: 10.27493/d.cnki.gzdzy.2021.000647.
[4] 李盼盼, 陈国旭, 刘盛东, 等. 基于高分卫星影像的断层构造识别及三维地质建模应用[J]. 合肥工业大学学报(自然科学版), 2020, 43(5): 680-687. DOI: 10.3969/j.issn.1003-5060.2020.05.018.
[5] VOLLGGER S A, CRUDEN A R. Mapping folds and fractures in basement and cover rocks using UAV photogrammetry, Cape Liptrap and Cape Paterson, Victoria, Australia[J]. Journal of Structural Geology, 2016, 85: 168-187. DOI: 10.1016/j.jsg.2016.02.012.
[6] FARAHBAKHSH E, CHANDRA R, OLIEROOK H K H, et al. Computer vision-based framework for extracting tectonic lineaments from optical remote sensing data[J]. International Journal of Remote Sensing, 2020, 41(5): 1760-1787. DOI: 10.1080/01431161.2019.1674462.
[7] 李宗仁, 张焜, 李晓民, 等. 遥感线性构造特征分析与找矿预测[J]. 遥感信息, 2016, 31(3): 115-121. DOI: 10.3969/j.issn.1000-3177.2016.03.018.
[8] 马艺璇, 李慧莉, 刘坤岩, 等. 基于分频相干体的蚂蚁追踪技术在塔河油田断裂刻画中的应用[J]. 石油物探, 2020, 59(2) : 258-266. DOI: 10.3969/j.issn.1000-1441.2020.02.012.
[9] 王怀秀, 冯思怡, 刘最亮. 基于改进随机森林算法的地质构造识别模型[J]. 煤炭科学技术, 2023, 51(4): 149-156. DOI: 10.13199/j.cnki.cst.2021-0754.
[10] 谢清惠, 蒋立伟, 赵春段, 等. 提高蚂蚁追踪裂缝预测精度的应用研究[J]. 物探与化探, 2021, 45(5): 1295-1302. DOI: 10.11720/wtyht.2021.1208.
[11] 杨楚龙, 王怀秀, 刘最亮. 基于改进XGBoost的地震多属性地质构造识别方法[J]. 科学技术与工程, 2023, 23(29): 12442-12450. DOI: 10.12404/j.issn.1671-1815.2023.23.29.12442.
[12] 孙振宇, 彭苏萍, 邹冠贵. 基于SVM算法的地震小断层自动识别[J]. 煤炭学报, 2017, 42(11): 2945-2952. DOI: 10.13225/j.cnki.jccs.2017.0972.
[13] 张荣, 李伟平, 莫同. 深度学习研究综述[J]. 信息与控制, 2018, 47(4): 385-397, 410. DOI: 10.13976/j.cnki.xk.2018.8091.
[14] ZHANG Y, WANG G, LI M C, et al. Automated classification analysis of geological structures based on images data and deep learning model[J]. Applied Sciences, 2018, 8(12): 2493. DOI: 10.3390/app8122493.
[15] 彭弘杰, 孙宇鹏, 赵禹深, 等. 基于Faster R-CNN对地质图进行褶皱自动识别[C] // 2021年中国地球科学联合学术年会论文集(十一): 专题三十一 深部矿产资源探测技术与应用、专题三十二 航空地球物理勘查技术与应用、专题三十三 煤炭资源与矿山地球物理. 北京: 北京伯通电子出版社, 2021: 1948-1950. DOI: 10.26914/c.cnkihy.2021.072331.
[16] 张黎, 吕芬, 尚凯, 等. 基于全卷积神经网络的走滑断裂识别技术: 以托甫台地区为例[J]. 长江大学学报(自然科学版), 2022, 19(4): 38-48. DOI: 10.16772/j.cnki.1673-1409.20220426.001.
[17] CUNHA A, POCHET A, LOPES H, et al. Seismic fault detection in real data using transfer learning from a convolutional neural network pre-trained with synthetic seismic data[J]. Computers & Geosciences, 2020, 135: 104344. DOI: 10.1016/j.cageo.2019.104344.
[18] 李辉. 基于改进AlexNet模型的断层识别方法[J]. 油气地质与采收率, 2022, 29(1): 107-112. DOI: 10.13673/j.cnki.cn37-1359/te.2022.01.013.
[19] 席英杰, 李克文, 徐延辉, 等. 一种用于地震断层图像识别的SPD-UNet模型[J]. 计算机工程, 2021, 47(12): 249-255. DOI: 10.19678/j.issn.1000-3428.0059327.
[20] 芦凤明, 孟瑞刚, 张军华, 等. UNet++和迁移学习相结合的复杂断裂识别方法研究[J]. 地球物理学进展, 2022, 37(3): 1100-1111. DOI: 10.6038/pg2022FF0122.
[21] TAN M X, LE Q. EfficientNet: rethinking model scaling for convolutional neural networks[C] // Proceedings of the 36th International Conference on Machine Learning. Long Beach, CA: PMLR, 2019: 6105-6114.
[22] HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C] // 2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Los Alamitos, CA: IEEE Computer Society, 2016: 770-778. DOI: 10.1109/CVPR.2016.90.
[23] ZAGORUYKO S, KOMODAKIS N. Wide residual networks[EB/OL]. (2017-06-14)[2024-01-09]. https://arxiv.org/abs/1605.07146. DOI: 10.48550/arXiv.1605.07146.
[24] HUANG Y P, CHENG Y L, BAPNA A, et al. GPipe: efficient training of giant neural networks using pipeline parallelism[C] // Advances in Neural Information Processing Systems 32 (NeurIPS 2019). Red Hook, NY: Curran Associates Inc., 2019: 103-112.
[25] 唐浪, 李慧霞, 颜晨倩, 等. 深度神经网络结构搜索综述[J]. 中国图象图形学报, 2021, 26(2): 245-264. DOI: 10.11834/jig.200202.
[26] DAI Z H, LIU H X, LE Q V, et al. CoAtNet: marrying convolution and attention for all data sizes[C] // Advances in Neural Information Processing Systems 34 (NeurIPS 2021). Red Hook, NY: Curran Associates Inc., 2021: 3965-3977.
[27] HUANG H J, CHEN Z G, ZOU Y, et al. Channel prior convolutional attention for medical image segmentation[EB/OL]. (2023-06-08)[2024-01-09]. https://arxiv.org/abs/2306.05196. DOI: 10.48550/arXiv.2306.05196.
[28] WANG Q L, WU B G, ZHU P F, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C] // 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2020: 11531-11539. DOI: 10.1109/CVPR42600.2020.01155.
[29] WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C] // Computer Vision-ECCV 2018: LNCS Volume 11211. Cham: Springer Nature Switzerland AG, 2018: 3-19. DOI: 10.1007/978-3-030-01234-2_1.
[30] SANTURKAR S, TSIPRAS D, ILYAS A, et al. How does batch normalization help optimization?[C] // Advances in Neural Information Processing Systems 31 (NeurIPS 2018). Red Hook, NY: Curran Associates Inc., 2018: 2488-2498.
[31] RAMACHANDRAN P, ZOPH B, LE Q V. Searching for activation functions[EB/OL]. (2017-10-27)[2024-01-09]. https://arxiv.org/abs/1710.05941. DOI: 10.48550/arXiv.1710.05941.
[32] GOODFELLOW I, BENGIO Y, COURVILLE A. Deep learning[M]. Cambridge, MA: MIT press, 2016.
[33] CHEN S W, SHIVAKUMAR S S, DCUNHA S, et al. Counting apples and oranges with deep learning: a data-driven approach[J]. IEEE Robotics and Automation Letters, 2017, 2(2): 781-788. DOI: 10.1109/LRA.2017.2651944.
[34] KAMILARIS A, PRENAFETA-BOLDÚ F X. Deep learning in agriculture: a survey[J]. Computers and Electronics in Agriculture, 2018, 147: 70-90. DOI: 10.1016/j.compag.2018.02.016.
[35] LONG M S, CAO Y, CAO Z J, et al. Transferable representation learning with deep adaptation networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, 41(12): 3071-3085. DOI: 10.1109/TPAMI.2018. 2868685.
[36] ZHUANG F Z, QI Z Y, DUAN K Y, et al. A comprehensive survey on transfer learning[J]. Proceedings of the IEEE, 2021, 109(1): 43-76. DOI: 10.1109/JPROC.2020.3004555.
[37] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[EB/OL]. (2015-04-10)[2024-01-09]. https://arxiv.org/abs/1409.1556. DOI: 10.48550/arXiv.1409.1556.
[38] HOWARD A, SANDLER M, CHEN B, et al. Searching for MobileNetV3[C] // 2019 IEEE/CVF International Conference on Computer Vision (ICCV). Los Alamitos, CA: IEEE Computer Society, 2019: 1314-1324. DOI: 10.1109/ICCV.2019. 00140.
[39] XIE S N, GIRSHICK R, DOLLÁR P, et al. Aggregated residual transformations for deep neural networks[C] // 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2017: 5987-5995. DOI: 10.1109/CVPR.2017.634.
[40] MA N N, ZHANG X Y, ZHENG H T, et al. ShuffleNet V2: practical guidelines for efficient CNN architecture design[C] //Computer Vision-ECCV 2018: LNCS Volume 11218. Cham: Springer Nature Switzerland AG, 2018: 122-138. DOI: 10.1007/978-3-030-01264-9_8.
[41] HUANG G, LIU Z, VAN DER MAATEN L, et al. Densely connected convolutional networks[C] // 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2017: 2261-2269. DOI: 10.1109/CVPR.2017.243.
[42] XU J, PAN Y, PAN X L, et al. RegNet: self-regulated network for image classification[J]. IEEE Transactions on Neural Networks and Learning Systems, 2023, 34(11): 9562-9567. DOI: 10.1109/tnnls.2022.3158966.
[43] MEHTA S, RASTEGARI M. MobileViT: light-weight, general-purpose, and mobile-friendly vision transformer[EB/OL]. (2022-03-04)[2024-01-09]. https://arxiv.org/abs/2110.02178. DOI: 10.48550/arXiv.2110.02178.
[44] LIU Z, MAO H Z, WU C Y, et al. A ConvNet for the 2020s[C] // 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2022: 11966-11976. DOI: 10.1109/CVPR52688.2022.01167.
[1] 戴林华, 黎远松, 石睿, 何忠良, 李雷. HSED-YOLO:一种轻量化的带钢表面缺陷检测模型[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 95-106.
[2] 李欣, 宁静. 基于时空特征融合的电力系统暂态稳定评估[J]. 广西师范大学学报（自然科学版）, 2024, 42(6): 89-100.
[3] 蒋懿波, 刘会家, 吴田. 基于改进残差网络的输电线路雷击过电压识别研究[J]. 广西师范大学学报（自然科学版）, 2023, 41(4): 74-83.
[4] 牛学德, 高丙朋, 任荣荣, 徐明明. 基于轻量级CNN的作物病虫害识别及安卓端应用[J]. 广西师范大学学报（自然科学版）, 2022, 40(6): 59-68.
[5] 田晟, 宋霖. 基于CNN和Bagging集成的交通标志识别[J]. 广西师范大学学报（自然科学版）, 2022, 40(4): 35-46.
[6] 帖军, 隆娟娟, 郑禄, 牛悦, 宋衍霖. 基于SK-EfficientNet的番茄叶片病害识别模型[J]. 广西师范大学学报（自然科学版）, 2022, 40(4): 104-114.
[7] 白捷, 高海力, 王永众, 杨来邦, 项晓航, 楼雄伟. 基于多路特征融合的Faster R-CNN与迁移学习的学生课堂行为检测[J]. 广西师范大学学报（自然科学版）, 2020, 38(5): 1-11.
[8] 严浩, 许洪波, 沈英汉, 程学旗. 开放式中文事件检测研究[J]. 广西师范大学学报（自然科学版）, 2020, 38(2): 64-71.
[9] 陈锦, 罗晓曙. 基于小波变换与野草算法的细胞图像特征提取与识别[J]. 广西师范大学学报（自然科学版）, 2015, 33(2): 22-28.
Viewed
Full text
0
HTML PDF
Just accepted Online first Issue Just accepted Online first Issue
0 0 0 0 0 0


Abstract
5
Just accepted Online first Issue
0 0 5
  From Others local
  Times 4 1
  Rate 80% 20%

Cited
Web of Science  Crossref   ScienceDirect  Search for Citations in Google Scholar >>
 
This page requires you have already subscribed to WoS.
  Shared   
  Discussed   
No Suggested Reading articles found!
版权所有 © 广西师范大学学报（自然科学版）编辑部
地址：广西桂林市三里店育才路15号　邮编：541004
电话：0773-5857325　E-mail: gxsdzkb@mailbox.gxnu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发