Improving Detection of a Portable NQR System for Humanitarian Demining Using Machine Learning

This article presents an approach to enhance the detection of buried landmines by applying machine learning (ML) to signals from a nuclear quadrupole resonance (NQR) system. This custom-made, low-cost, and portable NQR system has been developed for deployment in humanitarian demining, where strong radio frequency (RF) interference and a low signal-to-noise ratio (SNR) are important challenges for accurate detection. To tackle these problems, we have applied and tested various ML techniques to NQR signals acquired by our system in laboratory experiments with the explosive Research Department X (RDX). Results suggest that ML methods can indeed improve the detection accuracy of the NQR device, and this is confirmed further using data from field trials with our device. Importantly, the trained classifiers can be implemented with our device’s field-programmable gate array (FPGA) architecture and can run with little time penalty compared with simpler but less-efficient fast Fourier transform (FFT)-based energy detection methods.