Deep learning method for lithology classification of sandstone uranium deposits based on imaging spectroscopy

PAN Wei; ZHANG Yuantao; SHI Yuchen; CHEN Xuejiao

doi:10.12097/gbc.2023.12.007

2025 Vol. 44, No. 2~3

Article Contents

Article navigation > Geological Bulletin of China > 2025 > 44(2~3): 315-325

Next Article Previous Article

PAN Wei, ZHANG Yuantao, SHI Yuchen, CHEN Xuejiao. 2025. Deep learning method for lithology classification of sandstone uranium deposits based on imaging spectroscopy. Geological Bulletin of China, 44(2~3): 315-325. doi: 10.12097/gbc.2023.12.007

Citation:

PAN Wei, ZHANG Yuantao, SHI Yuchen, CHEN Xuejiao. 2025. Deep learning method for lithology classification of sandstone uranium deposits based on imaging spectroscopy. Geological Bulletin of China, 44(2~3): 315-325. doi: 10.12097/gbc.2023.12.007

Deep learning method for lithology classification of sandstone uranium deposits based on imaging spectroscopy

1.
National Key Laboratory of Remote Sensing Information and Image Analysis Technology, Beijing Research Institute of Uranium Geology, Beijing 100029, China
2.
No. 280 Institute of Nuclear Industry, Guanghan 618300, Sichuan, China

More Information

Corresponding author: ZHANG Yuantao, 809316163@qq.com

Received Date: 04 December 2023

Revised Date: 22 January 2024

Publish Date: 15 March 2025

Abstract

Abstract

Objective
Core logging is crucial for obtaining deep geological information about the earth. Currently, manual logging remains the primary method for acquiring lithological and other information, but it is time−consuming, labor−intensive, and prone to incomplete results and subjectivity.
Methods
Therefore, this study focuses on the ZKH3 of sandstone−type uranium deposits in the southwest of the Ordos Basin, applied deep learning and imaging spectroscopy techniques to core lithology identification. This study constructed a CNN model consisting of 7 one−dimensional convolutional layers, 2 pooling layers, 1 one−dimensional CBAM, and 3 fully connected layers.
Results
Additionally, a total of 26877 spectral samples from 7 types of rocks were collected, and model optimization and training were completed. Finally, the performance of the model was evaluated through comparison with Support Vector Machine (SVM) and its application in the whole borehole.The results show that the overall accuracy (OA) of deep learning model reached 94.6%, among which the producer’s accuracy (PA) of mudstone, fine sandstone, siltstone, medium sandstone, coarse sandstone, glutenite and background were 95.07%, 72.02%, 97.50%, 97.37%, 96.65%, 97.33%, and 99.01%, respectively. The Kappa coefficient was 0.94, which was better than SVM overall and achieved comparable results to geological logging.
Conclusions
This indicates that deep learning model based on imaging spectral data demonstrates excellent lithology classification and identification capabilities for core samples. And this approach enables non-destructive and rapid lithology identification while reducing the subjectivity inherent in manual geological logging to some extent, which provides valuable reference for digitization and automated logging research of core samples reference for research on digitalization and automation of core logging.
- imaging spectrum /
- uranium deposit /
- sandstone particle size /
- lithology classification /
- deep learning

FullText(HTML)

References

[1]	Agrawal N, Govil H. 2023. A deep residual convolutional neural network for mineral classification[J]. Advances in Space Research, 71(8): 3186−3202. doi: 10.1016/j.asr.2022.12.028 CrossRef Google Scholar
[2]	Alzubaidi F, Mostaghimi P, Swietojanski P, et al. 2021. Automated lithology classification from drill core images using convolutional neural networks[J]. Journal of Petroleum Science and Engineering, 197: 107933. doi: 10.1016/j.petrol.2020.107933 CrossRef Google Scholar
[3]	Baraboshkin E E, Ismailova L S, Orlov D M, et al. 2020. Deep convolutions for in−depth automated rock typing[J]. Computers & Geosciences, 135: 104330. Google Scholar
[4]	De La Rosa R, Khodadadzadeh M, Tusa L, et al. 2021. Mineral quantification at deposit scale using drill−core hyperspectral data: A case study in the Iberian Pyrite Belt[J]. Ore Geology Reviews, 139: 104514. doi: 10.1016/j.oregeorev.2021.104514 CrossRef Google Scholar
[5]	Ding S F, Qi B J, Tan H Y. 2011. An overview on theory and algorithm of support vector machines[J]. Journal of University of Electronic Science and Technology of China, 40(1): 2−10 (in Chinese with English abstract). Google Scholar
[6]	Ghanbari H, Antoniades D. 2022. Convolutional neural networks for mapping of lake sediment core particle size using hyperspectral imaging[J]. International Journal of Applied Earth Observation and Geoinformation, 112: 102906. doi: 10.1016/j.jag.2022.102906 CrossRef Google Scholar
[7]	Greenberger R N, Harris M, Ehlmann B L, et al. 2021. Hydrothermal alteration of the ocean crust and patterns in mineralization with depth as measured by micro‐imaging infrared spectroscopy[J]. Journal of Geophysical Research: Solid Earth, 126(8): e2021JB021976. doi: 10.1029/2021JB021976 CrossRef Google Scholar
[8]	Guan Y, Wang Q H, Feng J, et al. 2024. Comprehensive lithology recognition of altered igneous reservoirs based on machine learning for wireline and cutting logs in Huizhou depression, Pearl River Mouth Basin, northern South China Sea[J]. Journal of Jilin University (Earth Science Edition), 51(1): 345−358 (in Chinese with English abstract). Google Scholar
[9]	Guo T, Xu F, Ma J, et al. 2022. Component prediction of antai pills based on one−dimensional convolutional neural network and near−infrared spectroscopy[J]. Journal of Spectroscopy, (1): 1−10. Google Scholar
[10]	He K, Zhang X, Ren S, et al. 2016. Deep residual learning for image recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition: 770−778. Google Scholar
[11]	Houshmand N, GoodFellow S, Esmaeili K, et al. 2022. Rock type classification based on petrophysical, geochemical, and core imaging data using machine and deep learning techniques[J]. Applied Computing and Geosciences, 16: 100104. doi: 10.1016/j.acags.2022.100104 CrossRef Google Scholar
[12]	Hu W, Huang Y, Wei L, et al. 2015. Deep convolutional neural networks for hyperspectral image classification[J]. Journal of Sensors, 2015: 1−12. Google Scholar
[13]	Karimzadeh S, Tangestani M H. 2021. Evaluating the VNIR−SWIR datasets of WorldView−3 for lithological mapping of a metamorphic−igneous terrain using support vector machine algorithm: A case study of Central Iran[J]. Advances in Space Research, 68(6): 2421−2440. doi: 10.1016/j.asr.2021.05.002 CrossRef Google Scholar
[14]	Li P, Jiang N S, Feng Z P, et al. 2022. Freshly−opened swidden mapping using Support Vector Machine (SVM) and spatial characteristics in Phongsaly Province, Laos[J]. National Remote Sensing Bulletin, 26(11): 2329−2343 (in Chinese with English abstract). doi: 10.11834/jrs.20211113 CrossRef Google Scholar
[15]	Lu Y. 2018. Applications of hyperspectral mineral mapping in mineral and petroleum exploration[D]. China University of Geosciences (Beijing) PhD Thesis (in Chinese with English abstract). Google Scholar
[16]	Mathieu M, Roy R, Launeau P, et al. 2017. Alteration mapping on drill cores using a HySpex SWIR−320m hyperspectral camera: Application to the exploration of an unconformity−related uranium deposit (Saskatchewan, Canada)[J]. Journal of Geochemical Exploration, 172: 71−88. doi: 10.1016/j.gexplo.2016.09.008 CrossRef Google Scholar
[17]	Shirmard H, Farahbakhsh E, Müller R D, et al. 2022. A review of machine learning in processing remote sensing data for mineral exploration[J]. Remote Sensing of Environment, 268: 112750. doi: 10.1016/j.rse.2021.112750 CrossRef Google Scholar
[18]	Speta M, Gingras M K, Rivard B. 2016. Shortwave infrared hyperspectral imaging: A novel method for enhancing the visibility of sedimentary and biogenic features in oil−saturated core[J]. Journal of Sedimentary Research, 86(7): 830−842. doi: 10.2110/jsr.2016.54 CrossRef Google Scholar
[19]	Szegedy C, Vanhoucke V, Ioffe S, et al. 2016. Rethinking the inception architecture for computer vision[C]//Proceedings of the IEEE conference on computer vision and pattern recognition: 2818−2826. Google Scholar
[20]	Tappert M C, Rivard B, Fulop A, et al. 2015. Characterizing kimberlite dilution by crustal rocks at the Snap Lake diamond mine (Northwest Territories, Canada) using SWIR (1.90–2.36 μm) and LWIR (8.1–11.1 μm) hyperspectral imagery collected from drill core[J]. Economic Geology, 110(6): 1375−1387. doi: 10.2113/econgeo.110.6.1375 CrossRef Google Scholar
[21]	Tian Q L, Guo B J, Ye F W, et al. 2022. Mineral spectra classification based on one−dimensional dilated convolutional neural network[J]. Spectroscopy and Spectral Analysis, 42(3): 873−877 (in Chinese with English abstract). Google Scholar
[22]	Tusa L, Andreani L, Khodadadzadeh M, et al. 2019. Mineral mapping and vein detection in hyperspectral drill−core scans: Application to porphyry−type mineralization[J]. Minerals, 9(2): 122. doi: 10.3390/min9020122 CrossRef Google Scholar
[23]	Wan Y L, Zhong X W, Liu H, et al. 2021. Survey of application of convolutional neural network in classification of hyperspectral images[J]. Computer Engineering and Applications, 57(4): 1−10 (in Chinese with English abstract). Google Scholar
[24]	Woo S, Park J, Lee J Y, et al. 2018. Cbam: Convolutional block attention module[C]//Proceedings of the European conference on computer vision (ECCV), Germany: 3−19. Google Scholar
[25]	Ye B, Tian S, Cheng Q, et al. 2020. Application of lithological mapping based on advanced hyperspectral imager (AHSI) imagery onboard Gaofen−5 (GF−5) satellite[J]. Remote Sensing, 12(23): 3990. doi: 10.3390/rs12233990 CrossRef Google Scholar
[26]	Yu J, Zhang L, Li Q, et al. 2021. 3D autoencoder algorithm for lithological mapping using ZY−1 02D hyperspectral imagery: A case study of Liuyuan region[J]. Journal of Applied Remote Sensing, 15(4): 1−14. Google Scholar
[27]	Yu L, Porwal A, Holden E J, et al. 2012. Towards automatic lithological classification from remote sensing data using support vector machines[J]. Computers & Geosciences, 45: 229−239. Google Scholar
[28]	Zhang C, Ye F W, Yao J L, et al. 2017. Implementation method of drill core logging based on imaging spectrometer[J]. Remote Sensinng Information, 32(5): 69−74 (in Chinese with English abstract). Google Scholar
[29]	Zhang C, Ye F W, Xu Q J, et al. 2019. Deep drill logging and its alteration zoning features based on hyperspectral core imaging in west of Xiangshan uranium orefield[J]. Remote Sensing for Land and Resources, 31(2): 231−239 (in Chinese with English abstract). Google Scholar
[30]	Zhang C. 2020. Core imaging spectroscopy and three−dimensional alteration modeling for uranium deposit in Xiangshan, Jiangxi[D]. China University of Geosciences (Beijing), PhD Thesis: 1−150 (in Chinese with English abstract). Google Scholar
[31]	Zhang J L, Huang Y J, Wang J H, et al. 2013. Hyperspectral drilling core logging and 3D mineral mapping technology for uranium exploration[J]. Uranium Geology, 29(4): 249−255 (in Chinese with English abstract). Google Scholar
[32]	丁世飞, 齐丙娟, 谭红艳. 2011. 支持向量机理论与算法研究综述[J]. 电子科技大学学报, 40(1): 2−10. Google Scholar
[33]	管耀, 王清辉, 冯进, 等. 2024. 基于机器学习的蚀变火成岩测录井综合岩性识别——以南海北部珠江口盆地惠州26-6井区为例[J]. 吉林大学学报 (地球科学版), 54(1): 345−358. Google Scholar
[34]	李鹏, 蒋宁桑, 封志明, 等. 2022. 基于支持向量机的老挝丰沙里省新开辟刀耕火种遥感监测及其空间特征[J]. 遥感学报, 26(11): 2329−2343. Google Scholar
[35]	卢燕. 2018. 高光谱矿物填图技术在金属矿产和油气勘查中的应用研究[D]. 中国地质大学(北京) 博士学位论文. Google Scholar
[36]	田青林, 郭帮杰, 叶发旺, 等. 2022. 一维空洞卷积神经网络的矿物光谱分类[J]. 光谱学与光谱分析, 42(3): 873−877. doi: 10.3964/j.issn.1000-0593(2022)03-0873-05 CrossRef Google Scholar
[37]	万亚玲, 钟锡武, 刘慧, 等. 2021. 卷积神经网络在高光谱图像分类中的应用综述[J]. 计算机工程与应用, 57(4): 1−10. Google Scholar
[38]	张川. 2020. 岩心成像光谱技术与江西相山铀矿蚀变三维建模[D]. 中国地质大学(北京) 博士学位论文: 1−150. Google Scholar
[39]	张川, 叶发旺, 徐清俊, 等. 2019. 相山铀矿田西部深钻岩心成像光谱编录及蚀变分带特征[J]. 国土资源遥感, 31(2): 231−239. Google Scholar
[40]	张川, 叶发旺, 姚佳蕾, 等. 2017. 成像光谱钻孔岩芯编录的实现方法[J]. 遥感信息, 32(5): 69−74. Google Scholar
[41]	张杰林, 黄艳菊, 王俊虎, 等. 2013. 铀矿勘查钻孔岩心高光谱编录及三维矿物填图技术研究[J]. 铀矿地质, 29(4): 249−255. Google Scholar