Sparse Sensing and Machine Learning for Rapid Calibration of Defect Detection with rf Atomic Magnetometers - Paper

Aug 5, 2026

Machine learning algorithms are applied to accelerate both image acquisition and calibration for defect detection using radio-frequency atomic magnetometers. Sparse sensing is used to reconstruct images from a small fraction of sampled pixels, achieving an average relative error of about 5% compared with a full raster scan. Additional image-processing and calibration-support algorithms verify results, tune experimental parameters, and improve image contrast. The work establishes a first step toward a fast, arbitrary non-destructive defect-detection tool.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages23

Reading Time58 min

AccessFree to download

Introduction

Inductive measurement with rf atomic magnetometers

Magnetic induction tomography and penetration depth

Methods and approach

Sparse sensing for rapid calibration

Machine learning for image acquisition and reconstruction

Results and discussion

Reconstruction accuracy versus raster scanning

Image processing for qualification and contrast improvement

Edinburgh Research Explorer Sparse Sensing and Machine Learning for Rapid Calibration of Defect Detection with rf Atomic Magnetometers Citation for published version: Munko, M, Rushton, LM, Ellis, LM, Zipfel, JD, Bevington, P, Chalupczak, W & Lopez Dubon, S 2025,'Sparse Sensing and Machine Learning for Rapid Calibration of Defect Detection with rf Atomic Magnetometers', Sensors, vol. 25, no. 22, 6930. [https://doi.org/10.3390/s25226930](https://doi.org/10.3390/s25226930) Digital Object Identifier (DOI): 10.3390/s25226930 Link: Link to publication record in Edinburgh Research Explorer Document Version: Publisher's PDF, also known as Version of record Published In: Sensors General rights Copyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s) and / or other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights. Take down policy The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer content complies with UK legislation. If you believe that the public display of this file breaches copyright please [contact openaccess@ed.ac.uk](contact openaccess@ed.ac.uk) providing details, and we will remove access to the work immediately and investigate your claim. Download date: 08. Jan. 2026 Article Sparse Sensing and Machine Learning for Rapid Calibration of Defect Detection with rf Atomic Magnetometers Marek J. Munko 1, *, Lucas M. Rushton 2, Laura M. Ellis 2, Jake D. Zipfel 2, Patrick Bevington 2, *, Witold Chalupczak 2 and Sergio Lopez Dubon 1 Academic Editors: Yongfang Mao, Kunpeng Wang, Yonghua Jiang and Li Feng Received: 24 September 2025 Revised: 24 October 2025 Accepted: 31 October 2025 Published: 13 November 2025 Citation: Munko, M.J.; Rushton, L.M.; Ellis, L.M.; Zipfel, J.D.; Bevington, P.; Chalupczak, W.; Lopez Dubon, S. Sparse Sensing and Machine Learning for Rapid Calibration of Defect Detection with rf Atomic Magnetometers. Sensors 2025, 25, 6930. [https://doi.org/10.3390/](https://doi.org/10.3390/)[ ](https://doi.org/10.3390/)s25226930 Copyright: © 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://creativecommons.org/](https://creativecommons.org/)[ ](https://creativecommons.org/)[licenses/by/4.0/](licenses/by/4.0/)) . 1 School of Engineering, The University of Edinburgh, Mayfield Road, Edinburgh EH9 3JL, UK; [sergio.ldubon@ed.ac.uk](sergio.ldubon@ed.ac.uk) 2 National Physical Laboratory, Hampton Road, Teddington TW11 0LW, UK; [lucas.rushton@npl.co.uk](lucas.rushton@npl.co.uk) (L.M.R.); [laura.ellis@npl.co.uk](laura.ellis@npl.co.uk) (L.M.E.); [jake.zipfel@npl.co.uk](jake.zipfel@npl.co.uk) (J.D.Z.); [witold.chalupczak@npl.co.uk](witold.chalupczak@npl.co.uk) (W.C.) * [Correspondence: m.munko@ed.ac.uk](Correspondence: m.munko@ed.ac.uk) (M.J.M.); [patrick.bevington@npl.co.uk](patrick.bevington@npl.co.uk) (P.B.) Abstract Machine learning algorithms are utilised to improve the speed of both image acquisition and calibration processes needed for defect detection using radio-frequency atomic magnetometers. Sparse sensing is employed, and an average relative error of ≈5% is observed fora reconstructed image based on 1.25% of the sampled pixels when compared to a raster scan over the target object. Additional algorithms demonstrate the viability of image processing to qualify results and adjust experimental parameters required for calibration, leading to an enhancement of image contrast. This presents a first step in developing a tool for fast, arbitrary defect detection. Keywords: non-destructive testing; rf atomic magnetometers; sparse sensing; machine learning 1. Introduction There is a growing market need for rapid,

FAQs

How does sparse sensing reduce calibration time in defect detection?

Sparse sensing reconstructs defect images using only a small subset of sampled pixels, significantly reducing the number of measurements needed for calibration while maintaining acceptable accuracy.

What accuracy is reported for the reconstructed images?

The reported average relative error is approximately 5% when reconstructing an image from about 1.25% of sampled pixels, compared to a raster scan.

Why are radio-frequency atomic magnetometers suitable for detecting subsurface defects?

Oscillating magnetic fields penetrate with frequency-dependent depth, and inductive coupling reveals signals determined by the target’s electrical conductivity and magnetic permeability, enabling detection of concealed features.

Aug 5, 2026

Upload by Theodora