Scholarlink

<p>Lithological mapping is an effective tool for geological surveys and mineral exploration. However, it faces challenges in identifying complex rock types and improving classification accuracy. We mapped lithological units in the Karamaili ophiolite-mélange belt of Xinjiang using integrated machine learning algorithms, including artificial neural network (ANN), Mahalanobis distance (MD), support vector machine (SVM), and random forest (RF). These algorithms were utilized to process remote sensing datasets acquired by the Sustainable Development Science Satellite 1 Thermal Infrared Spectrometer (SDGSAT-1 TIS), Landsat-8 Operational Land Imager (OLI), and Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Digital Elevation Model (ASTER-GDEM). The results indicated that the overall accuracies of ANN, MD, SVM, and RF were 68.87%, 78.98%, 93.4%, and 98.36%, respectively. The SVM and RF effectively mapped the lithological units. The SDGSAT-1 TIS data helped to identify mafic–ultramafic and feldspar-rich rocks, while Landsat-8 OLI helped to successfully delineate granitoid and complex lithologies. The ASTER-GDEM data helped improve mapping accuracy by providing detailed topographic information. Thus, this study confirmed the efficacy of the implemented approaches to delineate mineralization zones and to discriminate lithological units. This study provides detailed geological data for lithological mapping and serves as a significant reference for geological surveys and environmental monitoring.</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

Lithologic Mapping in the Karamaili Ophiolite–Mélange Belt in Xinjiang, China, with Machine Learning and Integration of SDGSAT-1 TIS, Landsat-8 OLI and ASTER-GDEM

  • 期刊论文

  • 发表日期：

    2025年2月13日

查看全文
  • Zhao Zhang,
  • Fang Yin,
  • Yunqiang Zhu,
  • Lei Liu

摘要

Lithological mapping is an effective tool for geological surveys and mineral exploration. However, it faces challenges in identifying complex rock types and improving classification accuracy. We mapped lithological units in the Karamaili ophiolite-mélange belt of Xinjiang using integrated machine learning algorithms, including artificial neural network (ANN), Mahalanobis distance (MD), support vector machine (SVM), and random forest (RF). These algorithms were utilized to process remote sensing datasets acquired by the Sustainable Development Science Satellite 1 Thermal Infrared Spectrometer (SDGSAT-1 TIS), Landsat-8 Operational Land Imager (OLI), and Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Digital Elevation Model (ASTER-GDEM). The results indicated that the overall accuracies of ANN, MD, SVM, and RF were 68.87%, 78.98%, 93.4%, and 98.36%, respectively. The SVM and RF effectively mapped the lithological units. The SDGSAT-1 TIS data helped to identify mafic–ultramafic and feldspar-rich rocks, while Landsat-8 OLI helped to successfully delineate granitoid and complex lithologies. The ASTER-GDEM data helped improve mapping accuracy by providing detailed topographic information. Thus, this study confirmed the efficacy of the implemented approaches to delineate mineralization zones and to discriminate lithological units. This study provides detailed geological data for lithological mapping and serves as a significant reference for geological surveys and environmental monitoring.