VIIRS for noise detection

M-GEO
PLUS
STAMP
Topic description

Environmental noise has been associated with increasing blood pressure, heart disease, and sleep disturbance, among other complaints. Noise is traditionally modeled using complex computational programs. However, noise maps do not capture local issues, and many cities lack reliable noise information due to financial limitations. This is especially true in the global south.

The increasing availability of pre-processed remotely sensed data offers the opportunity to explore unconventional sources such as the VIIRS (Visible Infrared Imaging Radiometer Suite (VIIRS) for complex problems such as environmental noise mapping. The VIIRS instrument observes and collects global satellite observations that span the visible and infrared wavelengths across land, ocean, and atmosphere [1]. The Day/Night Band of VIIRS is being applied in a variety of applications, such as settlement detection and electric power grid detection. This research aims to answer a central research question:

To what extent can VIIRS support environmental noise detection and prediction?

Topic objectives and methodology

The research aims to uncover the relationship between modeled environmental noise and VIIRS data. In particular, you will implement an AI model, machine learning (ML) or deep learning (DL), to predict noise in a geographic area.

Data:

  • EO data, e.g., VIIRS time series, Sentinel-2 images, and elevation models
  • Modeled noise maps

Methods:

  • Machine learning methods: supervised, unsupervised, or their combination.
  • Collaborative planning workshop, maptables.

Workflow:

The suggested workflow of this research follows:

  1. Literature review of state-of-the-art ML/DL models for noise prediction using EO
  2. Data acquisition
  3. Implementation of the model
  4. Assessment of model performance
  5. Critical reflection on the model and its utility in consultation with key stakeholders.

This topic requires an interest in noise pollution and programming skills as you will work with Python. Also, there is ample room and opportunity to shape this research project in consultation with the supervisors, and we expect students who choose this topic to take the initiative to do so.

References for further reading
  • https://ladsweb.modaps.eosdis.nasa.gov/missions-and-measurements/viirs/
  • Linares Arroyo, H., Abascal, A., Degen, T., Aubé, M., Espey, B. R., Gyuk, G., Hölker, F., Jechow, A., Kuffer, M., Sánchez de Miguel, A., Simoneau, A., Walczak, K., & Kyba, C. C. M. (2024). Monitoring, trends and impacts of light pollution. Nature Reviews Earth & Environment5(6), 417-430. https://doi.org/10.1038/s43017-024-00555-9
  • J. Staab et al., "Using CNNs on Sentinel-2 data for road traffic noise modelling," 2023 Joint Urban Remote Sensing Event (JURSE), Heraklion, Greece, 2023, pp. 1-4, doi: 10.1109/JURSE57346.2023.10144160.
  • Comparison of land use regression and random forests models on estimating noise levels in five Canadian cities. Ying Liu et al. 2020. https://doi.org/10.1016/j.envpol.2019.113367
  • Aguilar, R., Flacke, J., Simon, D., & Pfeffer, K. (2023). Stakeholders Engagement in Noise Action Planning Mediated by OGITO: An Open Geo-Spatial Interactive Tool. Journal of Urban Technology, 30(3), 23–46. https://doi.org/10.1080/10630732.2023.2190705