Predicting air quality via multimodal AI and satellite imagery

Abstract

One of the most important major driving forces behind climate change and environmental issues that the Earth is currently facing is the pollution of the air which remains a key symptom of negative human influence on the environment. Air pollution is often invisible to the eye which can make its detection challenging, unlike the destruction of the land or waterways. Given that air-quality monitoring stations are typically ground-based, their abilities to detect pollutant distributions are often restricted to wide areas. Satellites, however, have the potential for studying the atmosphere at large; the European Space Agency (ESA) Copernicus project satellite, “Sentinel-5P” is a newly launched satellite capable of measuring a variety of pollutant information with publicly available data outputs. This paper seeks to create a multi-modal machine learning model for predicting air-quality metrics with high precision so that it will be applicable to locations where monitoring stations do not exist. The inputs of this model will include a fusion of ground measurements and satellite data with the goal of highlighting pollutant distribution and motivating change in societal and industrial behaviours. A contemporary method for fusing satellite information with pollution measurements is studied, suggesting that simpler models can work as effectively as neural network models that are constructed with state-of-the-art architectures. A new dataset of continental European pollution monitoring station measurements is created with features including altitude, population density, environmental classification of local areas, and satellite data from the ESA Copernicus project. This dataset is used to train a multi-modal ML model, Air Quality Network (AQNet) capable of fusing these various types of data sources to output predictions of various pollutants. These predictions are then aggregated to create an “air-quality index” that could be used to compare air quality over different regions. Three pollutants, NO2, O3, and PM10, are predicted successfully by AQNet and the network was found to be useful compared to a model only using satellite imagery. It was also found that the addition of supporting tabular data improves predictions. When testing the developed AQNet on out-of-sample data of the UK and Ireland, we obtain satisfactory estimates though on average pollution metrics were roughly overestimated by around 20%.