EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-SA 4.0

Image-based Air Quality Monitoring Using AI: Single-Image Estimation of AQI, PM2.5, and PM10.

Nora S. Salih, Mohammed A. Shakir

Zanin Journal of Science and Engineering · 30.03.2026 · 2 (1) · 47-65

DOI: 10.64362/zjse.172 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Zanin Journal of Science and Engineering, online erschienen am 30.03.2026. Verfasst von Nora S. Salih und 1 weiteren Autor:innen. Die Arbeit nennt 26 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Air pollution is a major threat to human health and the environment, so accessible monitoring is needed to support timely public action. Fixed monitoring stations are accurate but costly and sparsely deployed in many regions. In this study, an image-based deep learning approach was developed to estimate air quality from a single daytime outdoor photo. A compact multi-task model named ResSE-MTL (Residual + Squeeze–Excitation with Multi-Task Learning) was designed to take one RGB image (224×224×3) and predict AQI, PM2.5, and PM10; the numeric outputs were also mapped to six health levels for level-based reporting. Experiments were conducted on a public dataset from India and Nepal containing 12,240 daylight images labeled into six levels. On the held-out test set, R² = 0.99/0.99/0.97 and Accuracy = 90.24%/95.96%/95.42% for AQI/PM2.5/PM10. The model remained lightweight (~1.39M parameters), The main con-tribution is a lightweight multi-task image-based estimator that provides both numeric outputs and six-level reporting in a single forward pass, supported by a practical web interface. Limitations include reliance on a single dataset and restriction to daytime images; performance under night/low-light scenes and broader environmental and capture variability remains to be evaluated. Future work will extend coverage to more diverse conditions and include de-ployment-oriented evaluation on mobile/low-power devices.

Zitieren

APA

Salih, N. S., & Shakir, M. A. (2026). Image-based Air Quality Monitoring Using AI: Single-Image Estimation of AQI, PM2.5, and PM10. Zanin Journal of Science and Engineering, 2(1), 47-65. https://doi.org/10.64362/zjse.172

Vancouver

Salih NS, Shakir MA. Image-based Air Quality Monitoring Using AI: Single-Image Estimation of AQI, PM2.5, and PM10. Zanin Journal of Science and Engineering. 2026;2(1):47-65. doi:10.64362/zjse.172

BibTeX
@article{Salih2026,
  author = {Salih, Nora S. and Shakir, Mohammed A.},
  title = {Image-based Air Quality Monitoring Using AI: Single-Image Estimation of AQI, PM2.5, and PM10},
  journal = {Zanin Journal of Science and Engineering},
  year = {2026},
  volume = {2},
  number = {1},
  pages = {47--65},
  doi = {10.64362/zjse.172}
}
Kostenlos starten →