基于地质信息的改进GCN高光谱富锂铍伟晶岩信息提取方法
A Geological Information and Enhanced Graph Convolutional Network Method for Extracting Information on Lithium and Beryllium-Rich Pegmatites from Hyperspectral Imagery
近年来,随着高光谱卫星影像的发展,以及机器学习领域的技术突破,高光谱影像已在成矿预测领域取得众多成功应用.然而,传统的机器学习方法较多仅应用在高光谱数据上,往往忽视了地质成矿的复杂性,没有注意到地质领域信息对成矿的重要性.针对传统高光谱找矿中地质信息缺失的问题,本文将传统的高光谱数据与岩体、断层位置这类地质信息相结合,创建高光谱‒地质信息39通道综合数据集,同时对GCN(图卷积神经网络)模型进行改进,在网络中加入残差连接模块,同时对残差连接模块和卷积层进行批量化归一操作,加强训练效果.使用资源一号02D卫星(ZY-1 02D)高光谱数据对大红柳滩地区进行实验.结果表明,改进后的GCN模型对研究区内含矿花岗伟晶岩具有较高的识别精度.相比原始GCN网络、卷积神经网络模型和支持向量机模型,准确率分别提高了7、22和27个百分点,实现了高光谱遥感影像中锂铍矿化花岗伟晶岩的高精度自动化预测.
While advances in satellite hyperspectral technology and machine learning have significantly boosted its application in mineral prospectivity modeling, conventional data-driven approaches often fall short by neglecting the essential geological information that controls mineralization processes. To bridge the gap, this study develops a novel methodology that integrates hyperspectral imagery with critical geological determinants⁃specifically pluton boundaries and fault systems, establishing a 39-channel comprehensive data set featuring hyperspectral geological information, and introduces an enhanced Graph Convolutional Network (GCN) model. Architectural improvements include the incorporation of residual connections and the systematic application of batch normalization across both residual modules and convolutional layers, substantially stabilizing and accelerating the training process. Validation using ZY-1 02D hyperspectral data from the Dahongliutan area demonstrates that our refined GCN model achieves superior accuracy in identifying mineralized granitic pegmatites. Quantitative evaluations confirm substantial performance gains, with accuracy improvements of 7, 22, and 27 percentage points over the baseline GCN, Convolutional Neural Network, and Support Vector Machine models, respectively. This work establishes an effective and automated framework for high-precision prediction of lithium- and beryllium-mineralized granitic pegmatites via hyperspectral remote sensing.
| [1] |
Benson, T. R., Coble, M. A., Rytuba, J. J., et al., 2017. Lithium Enrichment in Intracontinental Rhyolite Magmas Leads to Li Deposits in Caldera Basins. Nature Communications, 8: 270. https://doi.org/10.1038/s41467⁃017⁃00234⁃y |
| [2] |
Cardoso⁃Fernandes, J., Teodoro, A. C., Lima, A., 2019. Remote Sensing Data in Lithium (Li) Exploration: A New Approach for the Detection of Li⁃Bearing Pegmatites. International Journal of Applied Earth Observation and Geoinformation, 76: 10-25. https://doi.org/10.1016/j.jag.2018.11.001 |
| [3] |
Dai, J. J., Wang, D. H., Dai, H. Z., et al., 2017. Geological Mapping and Ore⁃Prospecting Study Using Remote Sensing Technology in Jiajika Area of Western Sichuan Province. Geology in China. 44(2): 389-398 (in Chinese with English abstract). |
| [4] |
Ding, L., Chen, B. N., Zhu, Y. L., et al., 2024. Mineral Prediction Based on Prototype Learning. Computers & Geosciences, 184: 105540. https://doi.org/10.1016/j.cageo.2024.105540 |
| [5] |
Du, X. C., Lou, D. B., Xu, L. G., et al., 2023. Extracting Granite Pegmatite Information Based on GF⁃2 Images and the Random Forest Algorithm. Remote Sensing for Natural Resources, 35(4): 53-60 (in Chinese with English abstract). |
| [6] |
Fan, Y. H., Wang, H., Yang, X. K., et al., 2018. Application of High⁃Resolution Remote Sensing Technology to the Prospecting for Rare Metal Mineralization Belt. Remote Sensing for Natural Resources, 30(1): 128-134 (in Chinese with English abstract). |
| [7] |
Guan, Q. F., Ren, S. L., Chen, L. R., et al., 2022. Recognizing Multivariate Geochemical Anomalies Related to Mineralization by Using Deep Unsupervised Graph Learning. Natural Resources Research, 31(5): 2225-2245. https://doi.org/10.1007/s11053⁃022⁃10088⁃x |
| [8] |
Ioffe, S., Szegedy, C., 2015. Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. arXiv: 1502.03167. https://arxiv.org/abs/1502.03167 |
| [9] |
Jiang, G., Zhou, K. F., Wang, J. L., et al., 2023. Identification of Lithium⁃Beryllium Granitic Pegmatites Based on Deep Learning. Earth Science Frontiers, 30(5): 185-196 (in Chinese with English abstract). |
| [10] |
Jin, M. S., Gao, Y. B., Li, K., et al., 2019. Remote Sensing Prospecting Method for Pegmatite Type Rare Metal Deposit—Taking Dahongliutan Area in Western Kunlun for Example. Northwestern Geology, 52(4): 222-231 (in Chinese with English abstract). |
| [11] |
Kipf, T. N., Welling,M.,2017. Semi⁃Supervised Classification with Graph Convolutional Networks. arXiv, 1609.02907. https://doi.org/10.48550/arXiv.1609.02907 |
| [12] |
Liang, C. B., Xiao, B. H., Cheng, B., 2021. GCN⁃Based Semantic Segmentation Method for Mine Information Extraction in GAOFEN⁃1 Imagery. 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS., Brussels. https://doi.org/10.1109/IGARSS47720.2021.9554657 |
| [13] |
Lu, H., Qiao, D. Y., Li, Y. X., et al., 2021. Fusion of China ZY⁃1 02D Hyperspectral Data and Multispectral Data: Which Methods should be Used? Remote Sensing, 13(12): 2354. https://doi.org/10.3390/rs13122354 |
| [14] |
Pan, M., Tang, Y., Xiao, R. Q., et al., 2016. The Discovery of the Superlarge Li Ore Vein X03 in the Jiajika Ore District. Acta Geologica Sichuan, 36(3): 422-425, 430 (in Chinese with English abstract). |
| [15] |
Pathak, A. R., Pandey, M., Rautaray, S., 2018. Application of Deep Learning for Object Detection. Procedia Computer Science, 132: 1706-1717. https://doi.org/10.1016/j.procs.2018.05.144 |
| [16] |
Ren, G. L., Kong, H. L., Zhao, K. D., et al., 2022. Spectral Characteristics and Prospecting Implications of Lithium Deposits in Dahongliutan Area, Karakoram, Xinjiang. Northwestern Geology, 55(4): 103-114 (in Chinese with English abstract). |
| [17] |
Tu, Q. J., Han, Q., Li, P., et al., 2019a. Basic Characteristics and Exploration Progress of the Spodumene Ore Deposit in the Dahongliutan Area, West Kunlun. Acta Geologica Sinica, 93(11): 2862-2873 (in Chinese with English abstract). |
| [18] |
Tu, Q. J., Li, J. K., Wang, G., et al., 2019b. Mineralization Comparisons of the Major Pegmatite Type Spodumene Deposits and Their Prospecting Potential in West China. Geological Survey of China, 6(6): 35-47 (in Chinese with English abstract). |
| [19] |
Wang, H., Huang, L., Ma, H. D., et al., 2023. Geological Characteristics and Metallogenic Regularity of Lithium Deposits in Dahongliutan⁃Bailongshan Area, West Kunlun, China. Acta Petrologica Sinica, 39(7): 1931-1949 (in Chinese with English abstract). |
| [20] |
Wang, H. Y., 2021. Research on Pegmatite Dike Information Extraction from Remote Sensing Images Based on Deep Semantic Segmentation (Dissertation). China University of Geosciences, Beijing (in Chinese with English abstract). |
| [21] |
Wu, X. W., Sahoo, D., Hoi, S. C. H., 2020. Recent Advances in Deep Learning for Object Detection. Neurocomputing, 396: 39-64. https://doi.org/10.1016/j.neucom.2020.01.085 |
| [22] |
Xiong, X., Li, J. K., Yan, Q. G., 2024. The Ore⁃Forming Mechanism and Geological Indicators of the Zhawulong Pegmatite⁃Type Rare⁃Metal Deposit in Sichuan. Acta Petrologica Sinica, 40(9): 2863-2877 (in Chinese with English abstract). |
| [23] |
Zhang, S., Ju, N., Wu, Y., et al., 2023. Distribution Characteristics, Main Types and Exploration and Development Status of Beryllium Deposit. Geology in China, 50(2): 410-424 (in Chinese with English abstract). |
| [24] |
Zhao, Z. Q., Zheng, P., Xu, S. T., et al., 2019. Object Detection with Deep Learning: A Review. IEEE Transactions on Neural Networks and Learning Systems, 30(11): 3212-3232. https://doi.org/10.1109/TNNLS.2018.2876865 |
| [25] |
Zuo, R. G., Xu, Y., 2023. Graph Deep Learning Model for Mapping Mineral Prospectivity. Mathematical Geosciences, 55(1): 1-21. https://doi.org/10.1007/s11004⁃022⁃10015⁃z |
| [26] |
Zuo, R. G., Xu, Y., 2024. A Physically Constrained Hybrid Deep Learning Model to Mine a Geochemical Data Cube in Support of Mineral Exploration. Computers & Geosciences, 182: 105490. https://doi.org/10.1016/j.cageo.2023.105490 |
国家自然科学基金重点项目(42430111)
/
| 〈 |
|
〉 |