Machine learning and remote sensing integration for leveraging urban sustainability: A review and framework

• Published in: Sustainable cities and society 2023-09, Vol.96, p.104653, Article 104653
• Main Authors: Li, Fei, Yigitcanlar, Tan, Nepal, Madhav, Nguyen, Kien, Dur, Fatih
• Format: Article
• Language: English
• Subjects: Deep learning; Machine learning; Remote sensing; Sustainable urban development; Urban analytics; Urban studies; Urban sustainability
• ISSN: 2210-6707
• DOI: 10.1016/j.scs.2023.104653

•Insights into machine learning and remote sensing integration in urban studies are generated.•Indicators of physical attributes over socioeconomic factors are widely utilised.•Conventional satellite, aerial images, and Lidar data are prevalent due to easy accessibility.•Integration offers accurate results and thorough analysis in image processing and analytics.•Image acquisition and decision-making necessitate human supervision for accurate outputs. Climate change and rapid urbanisation exacerbated multiple urban issues threatening urban sustainability. Numerous studies integrated machine learning and remote sensing to monitor urban issues and develop mitigation strategies for sustainability. However, few studies comparatively analysed joint applications of machine learning and remote sensing for urban issues and sustainability. This paper presents a systematic review and formulates a framework integrating machine learning and remote sensing in urban studies. The literature analysis reveals: Most studies occurred in Asia, Europe, and North America, driven by technical and ethical factors, highlighting responsible approaches for data-scarce regions; Reviewed studies prioritised physical spatial aspects over socioeconomic factors, requiring multi-source data for comprehensive analysis; Conventional satellite, aerial images, and Lidar data are prevalent due to affordability, quality, and accessibility; Although supervised machine learning dominates, unsupervised methods and algorithm selection paradigms require exploration; Integration offers accurate results and thorough analysis in image processing and analytics, while image acquisition and decision-making necessitate human supervision. This paper provides a comprehensive review and an integrative framework for machine learning and remote sensing, enriching insights into their potential for urban studies and spatial analytics. The study informs urban planning and policymaking by promoting efficient management via enhanced machine learning and remote sensing integration, and bolstering data-driven decision-making.