Deep Transfer Learning on Satellite Imagery Improves Air Quality Estimates in Developing Nations

TL;DR

This paper presents a transfer learning approach using satellite imagery to estimate air quality in developing countries, addressing data scarcity and improving public health risk assessments.

Contribution

It introduces a scalable deep transfer learning method that adapts models trained in high-income countries to low- and middle-income city contexts.

Findings

Effective AQ estimation in Accra using transfer learning from US cities

Demonstrates the potential of satellite imagery for AQ monitoring in data-scarce regions

Improves emergency preparedness and risk mitigation in LMICs

Abstract

Urban air pollution is a public health challenge in low- and middle-income countries (LMICs). However, LMICs lack adequate air quality (AQ) monitoring infrastructure. A persistent challenge has been our inability to estimate AQ accurately in LMIC cities, which hinders emergency preparedness and risk mitigation. Deep learning-based models that map satellite imagery to AQ can be built for high-income countries (HICs) with adequate ground data. Here we demonstrate that a scalable approach that adapts deep transfer learning on satellite imagery for AQ can extract meaningful estimates and insights in LMIC cities based on spatiotemporal patterns learned in HIC cities. The approach is demonstrated for Accra in Ghana, Africa, with AQ patterns learned from two US cities, specifically Los Angeles and New York.