Advances in electromagnetic techniques for subsurface infrastructure detection: A comprehensive review of methods, challenges, and innovations

TL;DR

This comprehensive review examines electromagnetic and data fusion techniques for detecting buried infrastructure, highlighting their advantages, limitations, and future research directions to improve accuracy and real-time processing.

Contribution

It provides an integrated overview of current methods, challenges, and innovations in electromagnetic subsurface detection, emphasizing data fusion and mathematical estimators.

Findings

Data fusion improves detection accuracy.

Magnetometry effectively detects ferrous targets.

Mathematical estimators reduce noise and enhance survey precision.

Abstract

This review paper explores the state-of-the-art in non-intrusive methods for detecting and characterising buried infrastructure, focusing on Electrical Resistivity Tomography (ERT), Infrared Thermography (IRT), and magnetometry, along with data fusion techniques and mathematical estimators. ERT and IRT offer distinct advantages in subsurface imaging, while magnetometry provides omnidirectional measurements ideal for detecting ferrous targets. Despite these benefits, each method has inherent limitations, such as challenges in depth estimation and difficulties in distinguishing between various subsurface objects. The integration of multiple sensing techniques through data fusion approaches has shown significant promise in overcoming these limitations and improving detection accuracy. Additionally, mathematical estimators, including Kalman filters and particle filters, play a crucial role in reducing noise and enhancing the precision of geophysical surveys. This review discusses the strengths, limitations, and future research needs of these techniques, offering a comprehensive understanding of their current and potential applications in buried infrastructure detection. The paper concludes by emphasising the importance of optimising sensor performance, refining fusion algorithms, and exploring hybrid models for real-time data processing in future research.