Scholarlink

<p>The rapid expansion of geological exploration data has highlighted the limitations of traditional mineral resource assessment methods and increased the need for advanced data fusion approaches that can integrate heterogeneous datasets. This study evaluated two multi-source data fusion models for assessing both shallow and deep-seated Pb–Zn mineralization at the Fankou deposit (northern Guangdong, China) and its surrounding area. Surface mineralization indicators were extracted from remote sensing and geophysical datasets, and these indices were integrated using the ordered weighted averaging and fuzzy-analytic hierarchical process methods, yielding receiver operating characteristic–area under the curve values of 0.764 and 0.938, respectively. To further enhance surface prospectivity mapping, a semantic-segmentation fully convolutional network model was developed, achieving an overall accuracy of 0.91. Guided by the identified zones of mineral prospectivity, the Fankou mine, which exhibits strong prospectivity for Pb–Zn mineralization, was selected for in-depth evaluation. In total, 51 well-logging datasets, 32 controlled-source audio-frequency magnetotelluric profiles, three radon measurements, two electrochemical datasets, and magnetometry data were analyzed to characterize subsurface geological features. These datasets were integrated in GeoModeller to construct a three-dimensional lithological model, which was subsequently used to perform a litho-constrained 3D inversion to pinpoint prospective mineralization zones at depth. Structural analysis indicates that mineralization is primarily controlled by NNE-, N–S-, and E–W-trending fractures. Overall, this study presents a novel methodology for targeting Pb–Zn ore, encompassing both shallow and deep-seated deposits by integrating multi-source geological data, thereby improving the evaluation of polymetallic resources in complex geological settings.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-Dimensional Assessment of the Fankou Pb–Zn Deposit, China: A Combined Data Fusion and Deep Learning Approach

  • 期刊论文

  • 发表日期：

    2026年3月19日

查看全文
  • Joel Paterne Kouame,
  • Jianguo Chen,
  • Chengbin Wang,
  • Abara A. Biabak Indrik,
  • Beugre Kossou Yves Bertrand,
  • Guangjun Ji

摘要

The rapid expansion of geological exploration data has highlighted the limitations of traditional mineral resource assessment methods and increased the need for advanced data fusion approaches that can integrate heterogeneous datasets. This study evaluated two multi-source data fusion models for assessing both shallow and deep-seated Pb–Zn mineralization at the Fankou deposit (northern Guangdong, China) and its surrounding area. Surface mineralization indicators were extracted from remote sensing and geophysical datasets, and these indices were integrated using the ordered weighted averaging and fuzzy-analytic hierarchical process methods, yielding receiver operating characteristic–area under the curve values of 0.764 and 0.938, respectively. To further enhance surface prospectivity mapping, a semantic-segmentation fully convolutional network model was developed, achieving an overall accuracy of 0.91. Guided by the identified zones of mineral prospectivity, the Fankou mine, which exhibits strong prospectivity for Pb–Zn mineralization, was selected for in-depth evaluation. In total, 51 well-logging datasets, 32 controlled-source audio-frequency magnetotelluric profiles, three radon measurements, two electrochemical datasets, and magnetometry data were analyzed to characterize subsurface geological features. These datasets were integrated in GeoModeller to construct a three-dimensional lithological model, which was subsequently used to perform a litho-constrained 3D inversion to pinpoint prospective mineralization zones at depth. Structural analysis indicates that mineralization is primarily controlled by NNE-, N–S-, and E–W-trending fractures. Overall, this study presents a novel methodology for targeting Pb–Zn ore, encompassing both shallow and deep-seated deposits by integrating multi-source geological data, thereby improving the evaluation of polymetallic resources in complex geological settings.