Back to search

University of Southampton

Assessing the feasibility of applying machine learning tools to predict environmental stressors from digital urban fingerprints

Abstract

dc:description.abstract

Urban noise pollution poses persistent challenges to public health and urban sustainability. This dissertation advances a new technical paradigm for scalable urban noise prediction by integrating multispectral remote sensing imagery, land use/land cover (LULC) data, and state-of-the-art machine learning techniques. The research adopts a progressive structure comprising three data-driven studies, each representing a major methodological step forward.<br/>Chapter 4 presents a pioneering approach that leverages convolutional neural networks (CNNs) to predict citywide noise levels using high-resolution multispectral imagery, validated in Southampton. Chapter 5 incorporates geospatial relationships through graph-based modeling, further improving spatial prediction accuracy. Building on these foundations, Chapter 6 proposes a generalizable dual-branch graph neural network (GNN) framework with domain adaptation and pseudo-labeling, enabling robust noise mapping across five UK cities using standardized remote sensing and Urban Atlas LULC data.<br/>Results demonstrate that deep learning models—when properly integrated with remote sensing and urban structural features—can achieve high accuracy and transferability in noise prediction, even in cities lacking extensive noise measurements. The workflow substantially reduces field data collection costs and advances urban noise assessment toward scalable, transferable solutions.<br/>This dissertation thus bridges the gap between traditional acoustic modeling and next-generation data-driven mapping, providing methodological innovations with practical value for urban planners and environmental authorities seeking efficient, city-scale noise management tools.

Degree

thesis:*
Name dc:type.qualificationname
Ph.D.
Level dc:type.qualificationlevel
doctoral
Grantor dc:publisher.institution
University of Southampton
Year dc:date.issued
2025

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Zhu, Feiyu
Advisors dc:contributor.advisor
  • Eigenbrod, Felix
  • Hare, Jonathon

Rights

Language dc:language
English, eng

Chain of custody

source
Harvested from
University of Southampton
Base URL
eprints.soton.ac.uk/cgi/oai2
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
related terms
citation

Zhu, Feiyu. Assessing the feasibility of applying machine learning tools to predict environmental stressors from digital urban fingerprints. doctoral thesis, University of Southampton, 2025.