EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Machine-Learning-Enhanced Transient Electromagnetic Signal Processing for Advanced Mineral Exploration

Changchun Wang, Maohong You, Fei Qin

Journal of Environmental & Earth Sciences · 17.04.2026 · 157-183

DOI: 10.30564/jees.v8i4.13043 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Environmental & Earth Sciences, online erschienen am 17.04.2026. Verfasst von Changchun Wang und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This review synthesizes recent progress in machine-learning-enhanced transient electromagnetic signal processing for advanced mineral exploration, with emphasis on methods that improve reliability under realistic field conditions. Transient Electromagnetic (TEM) data provide strong sensitivity to subsurface conductivity, yet practical interpretation is often limited by non-stationary interference, platform- and instrument-dependent artifacts, wide dynamic range decay, and ill-posed inversion. We organize the literature using a pipeline perspective spanning automated quality control, denoising and interference suppression, system-response correction and normalization, representation learning, Machine learning (ML) assisted inversion (including surrogate forward models and hybrid physics ML inference), and target detection and ranking. Particular attention is given to the exploration-specific constraints that shape evidence quality, including label scarcity and bias, the synthetic-to-field gap, spatial leakage in evaluation splits, and the need for cost-aware metrics tied to drill decisions rather than pointwise regression error. Across reported studies, the strongest and most transferable benefits are observed in quality control (QC) automation and interference-aware denoising that improve repeatability and stabilize downstream inversion. More ambitious end-to-end inversion and target-ranking models remain promising but are highly sensitive to domain shift across waveforms, gate schedules, noise regimes, and geology, making calibrated uncertainty estimation and out-of-distribution detection central requirements for deployment. We conclude by outlining reproducibility and reporting practices suitable for SCI-standard evidence and by identifying priority research directions, including realistic synthetic data generation, hybrid inversion with error control, and benchmark tasks aligned with exploration value.

Zitieren

APA

Wang, C., You, M., & Qin, F. (2026). Machine-Learning-Enhanced Transient Electromagnetic Signal Processing for Advanced Mineral Exploration. Journal of Environmental & Earth Sciences, 157-183. https://doi.org/10.30564/jees.v8i4.13043

Vancouver

Wang C, You M, Qin F. Machine-Learning-Enhanced Transient Electromagnetic Signal Processing for Advanced Mineral Exploration. Journal of Environmental & Earth Sciences. 2026:157-183. doi:10.30564/jees.v8i4.13043

BibTeX
@article{Wang2026,
  author = {Wang, Changchun and You, Maohong and Qin, Fei},
  title = {Machine-Learning-Enhanced Transient Electromagnetic Signal Processing for Advanced Mineral Exploration},
  journal = {Journal of Environmental & Earth Sciences},
  year = {2026},
  pages = {157--183},
  doi = {10.30564/jees.v8i4.13043}
}
Kostenlos starten →