Various techniques are used to assess air quality. Basic parameters such as particulate matter (PM) and certain gases can be easily obtained from local meteorological stations. However, for more detailed data, such as heavy metal concentrations, researchers must collect and analyze samples in laboratories. Due to natural limitations, regulatory monitoring is often restricted in both spatial and temporal coverage.

Satellite imagery provides a valuable source of atmospheric and surface data. Each year, new missions with advanced sensors enhance remote sensing capabilities. Modern instruments like Sentinel-5 offer near-ready air quality data, including information on gases and aerosols. However, the Sentinel-5's orbital cycle and resolution remain limited. Meanwhile, widely used public satellite missions such as Landsat, MODIS, and Sentinel provide high-resolution data with frequent updates. Integrating in situ measurements with satellite data and machine learning techniques enhances air quality monitoring. Modeling helps fill gaps in in situ data, provides detailed assessments of specific areas, and enables a partial automation of environmental control processes.

This report reviews widely used satellite-based programs, tools for efficient data processing, and machine learning approaches for air quality estimation. It highlights the effectiveness and advantages of ML-driven remote sensing for air quality monitoring. Additionally, we discuss commercial satellite missions, firsthand experiences, and future directions for advancing air quality monitoring technologies.