A Low-Cost Image Sensor for Particulate Matter Detection to Streamline Citizen Science Campaigns on Air Quality Monitoring

Air pollution poses a significant threat to public health by contributing to respiratory and chronic diseases. While effective monitoring of
$\boldsymbol{PM}_{10}$
particles is crucial for managing this issue, existing low-cost methods require enhancement for broader community involvement and accuracy. This paper shows an image processing technique for detecting and quantifying
$\boldsymbol{PM}_{10}$
particles with Do-it-Yourself (DIY) sensors that incorporates citizen science, allowing community and student participation in data collection. Comparing four clustering algorithms-Spectral, BIRCH, Gaussian Mixture, and K Means-the method proposed demonstrates superior performance with K Means in processing both synthetic and real-world images. The initial tests on synthetic datasets established baseline metrics, which were therefore validated through field data from Bilbao, Spain, collected by local university students. This approach not only improved the detection and quantification of
$\boldsymbol{PM}_{10}$
particles using an extremely low-cost sensor but also enhanced public engagement and awareness of environmental issues. The findings suggest that integrating citizen science with advanced image processing can significantly advance air quality monitoring and promote democratisation of environmental sensing.