A review of machine learning in processing remote sensing data for mineral exploration (2103.07678v2)

Abstract: The decline of the number of newly discovered mineral deposits and increase in demand for different minerals in recent years has led exploration geologists to look for more efficient and innovative methods for processing different data types at each stage of mineral exploration. As a primary step, various features, such as lithological units, alteration types, structures, and indicator minerals, are mapped to aid decision-making in targeting ore deposits. Different types of remote sensing datasets, such as satellite and airborne data, make it possible to overcome common problems associated with mapping geological features. The rapid increase in the volume of remote sensing data obtained from different platforms has encouraged scientists to develop advanced, innovative, and robust data processing methodologies. Machine learning methods can help process a wide range of remote sensing datasets and determine the relationship between components such as the reflectance continuum and features of interest. These methods are robust in processing spectral and ground truth measurements against noise and uncertainties. In recent years, many studies have been carried out by supplementing geological surveys with remote sensing datasets, which is now prominent in geoscience research. This paper provides a comprehensive review of the implementation and adaptation of some popular and recently established machine learning methods for processing different types of remote sensing data and investigates their applications for detecting various ore deposit types. We demonstrate the high capability of combining remote sensing data and machine learning methods for mapping different geological features that are critical for providing potential maps. Moreover, we find there is scope for advanced methods to process the new generation of remote sensing data for creating improved mineral prospectivity maps.

• The paper reviews diverse ML algorithms such as PCA, SVM, and deep learning for processing remote sensing data to enhance mineral exploration.
• The paper details how clustering, regression, and CNN techniques improve geological mapping and aid in accurately identifying mineral deposits.
• The paper discusses challenges like integrating multi-source datasets and addressing class imbalance while advocating for advanced deep learning in future research.

A Review of Machine Learning in Processing Remote Sensing Data for Mineral Exploration

The paper "A review of machine learning in processing remote sensing data for mineral exploration" by Shirmard et al. explores the intersection of ML methodologies and remote sensing data within the context of mineral exploration. This work attempts to address the ever-increasing challenges in mineral exploration, primarily driven by declining mineral discoveries and escalating mineral demand.

The authors provide a comprehensive survey of remote sensing technologies used in geological studies, detailing various types of satellite, airborne, and ground-based data acquisition methods. The text underscores how these diverse datasets, ranging from optical to radar data, are employed for geological mapping. The discussion extends to specific data-acquisition systems, such as ASTER, Landsat, and Sentinel satellites, exploring their band compositions and resolutions, alongside the LiDAR and drone-borne systems used for high-precision surveying.

The core contribution of this paper is its extensive examination of ML algorithms applied to process these remote sensing datasets, in service of mineral exploration. The paper categorizes ML methodologies into several types:

• Dimensionality Reduction: Techniques like Principal Component Analysis (PCA), Independent Component Analysis (ICA), and Minimum Noise Fraction (MNF) are leveraged to distill complex multispectral data into more manageable forms. These approaches, particularly PCA, have been effectively used to identify lithological units and alteration zones related to mineralization.
• Classification: Methods such as Support Vector Machines (SVMs), Random Forests, and Convolutional Neural Networks (CNNs) play prominent roles. SVMs and Random Forests have demonstrated efficacy in lithological mapping and the classification of mineral deposits, whereas CNNs, with their hierarchical feature learning capability, offer significant promise for automatic feature extraction from remote sensing data.
• Clustering: Algorithms like K-means and ISODATA facilitate the grouping of spectral data points into clusters, assisting in identifying mineralogical compositions.
• Regression: Regression approaches enable predictive modeling of mineral presence over various geological settings.
• Deep Learning: With the advent of deep learning, traditional ML has undergone a paradigm shift, introducing architectures capable of modeling complex, non-linear relationships within large datasets. Deep learning methods, specifically LSTMs and hybrid models like LSTM-CNN, are explored for their potential in remote sensing applications.

The paper emphasizes potential breakthroughs achievable through advanced ML techniques, such as Graph Convolutional Networks (GCNs) for spatial data and Bayesian Neural Networks for uncertainty quantification. Despite the evident advantages of ML in enhancing remote sensing data interpretation, the authors recognize challenges, including the integration of multi-source datasets and dealing with class imbalance in training data—a common issue in mineral exploration studies.

The implications of this research extend beyond theoretical insights, proposing practical benefits in the discovery and mapping of mineral deposits, essential for meeting the increasing global demands for raw materials. Future directions pointed to involve refining deep learning methodologies for higher accuracy and efficiency, leveraging advancements in big data analytics, and continued cross-disciplinary collaboration.

This in-depth review reinforces the importance of machine learning's evolving role in the geosciences, offering a roadmap for the utilization of these methodologies in mineral exploration and positioning itself as a reference for ongoing and future research endeavors in the field.