Earth's Field Magnetic Resonance Imaging in Urban Environments Noise Suppression and Advanced Image Reconstruction Techniques

Abstract

This dissertation investigates the feasibility and practical implementation of Earth’s Field Magnetic Resonance Imaging (EF-MRI) under real-world urban noise environments, with experiments conducted at Trinity College Dublin (city centre) and a Dublin suburb. Oper- ating at the Earth’s magnetic field ( 50 μT) with ultra-low Larmor frequencies ( 1.9 kHz), EF-MRI offers a low-cost, portable alternative to conventional high-field MRI systems, addressing accessibility challenges in resource-limited settings. However, environmental noise, electromagnetic interference, low polarization, and frequency drift remain critical challenges for its practical deployment.

Using the Magritek Terranova EFNMR system, this work conducted imaging experi- ments on two distinct water bottle phantom samples, employing both the Filtered Back Projection (FBP) method and k-space imaging techniques to evaluate reconstruction per- formance under different noise environments. To mitigate the impact of environmental noise, the study integrates systematic calibration, including AutoShim for field homogeneity optimization, B1 pulse duration adjustment, and environmental noise mapping, combined with advanced signal processing approaches such as exponential temporal weighting, frequency correction, and k-space filtering.

1D, 2D, and 3D EF-MRI imaging experiments demonstrate successful noise suppres- sion and improved image quality in unshielded environments, achieving enhanced Signal- to-Noise Ratio (SNR) and Structural Similarity Index (SSIM). Comparative evaluations with default Prospa reconstructions further confirm the proposed pipeline’s effectiveness in improving spatial resolution, contrast, and structural fidelity. This work demonstrates that combining system optimization with advanced signal processing enables stable and practical EF-MRI imaging under urban conditions, providing experimental evidence and methodological insights for the future deployment of low-cost MRI in community-level healthcare.