Scholarlink

<p>Airborne geophysical data are widely used for data integration and in mineral prospectivity modeling. In this study, airborne geophysical data were used to create a Pseudo-geological map in the Shahr-e-Babak study area in southeastern Iran. The K-means, fuzzy c-means (FCM), and self-organized map (SOM) unsupervised machine learning methods were used to cluster airborne geophysics data. All data were normalized before being input into the machine learning algorithms. All normalized data were input to the unsupervised algorithm method. The Silhouette coefficient, Calinski-Harabasz index, and Davies-Bouldin index were used to determine the optimal number of clusters. Post-clustering evaluation was conducted using Random Forest (RF), SOM, and Silhouette coefficient. The clustering evaluation results indicate that SOM is a more suitable method for clustering airborne geophysical data compared to the other two methods. After creating a pseudo-geological map, the resulting map was correlated with the geological map of the study area. Cluster 1 corresponds to trachyandesite, trachybasalt, and granodiorite; cluster 2 to volcanic rocks, flysch, and conglomerate; cluster 3 to the volcanic rocks and dacite rocks; cluster 4 to the andesite and basalt lithologies; cluster 5 to the pyroclastic units and quaternary units, and finally, cluster 6 to the quaternary. The generated pseudo-geological map can also be used to create a mineral potential model. The mineral potential model generated by the SOM method is preferable to the other two methods due to the area of the possible area and the identification of the number of indices. The high potential areas for copper mineralization were suggested, and most of these areas are located in the northwestern and central parts of the study area. In this paper, a general framework for preprocessing and processing airborne geophysical data, as well as the application of unsupervised machine learning methods and the selection of a more optimal method, is proposed for creating a pseudo-geological map of the Shahr-e-Babak study area.</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

Evaluating unsupervised machine learning techniques for geological mapping with airborne geophysical data: case study of Shahr-e-Babak, Iran

  • 期刊论文

  • 发表日期：

    2025年12月10日

查看全文
  • Moslem Jahantigh,
  • Hamidreza Ramazi

摘要

Airborne geophysical data are widely used for data integration and in mineral prospectivity modeling. In this study, airborne geophysical data were used to create a Pseudo-geological map in the Shahr-e-Babak study area in southeastern Iran. The K-means, fuzzy c-means (FCM), and self-organized map (SOM) unsupervised machine learning methods were used to cluster airborne geophysics data. All data were normalized before being input into the machine learning algorithms. All normalized data were input to the unsupervised algorithm method. The Silhouette coefficient, Calinski-Harabasz index, and Davies-Bouldin index were used to determine the optimal number of clusters. Post-clustering evaluation was conducted using Random Forest (RF), SOM, and Silhouette coefficient. The clustering evaluation results indicate that SOM is a more suitable method for clustering airborne geophysical data compared to the other two methods. After creating a pseudo-geological map, the resulting map was correlated with the geological map of the study area. Cluster 1 corresponds to trachyandesite, trachybasalt, and granodiorite; cluster 2 to volcanic rocks, flysch, and conglomerate; cluster 3 to the volcanic rocks and dacite rocks; cluster 4 to the andesite and basalt lithologies; cluster 5 to the pyroclastic units and quaternary units, and finally, cluster 6 to the quaternary. The generated pseudo-geological map can also be used to create a mineral potential model. The mineral potential model generated by the SOM method is preferable to the other two methods due to the area of the possible area and the identification of the number of indices. The high potential areas for copper mineralization were suggested, and most of these areas are located in the northwestern and central parts of the study area. In this paper, a general framework for preprocessing and processing airborne geophysical data, as well as the application of unsupervised machine learning methods and the selection of a more optimal method, is proposed for creating a pseudo-geological map of the Shahr-e-Babak study area.