Deep learning-based pollen imaging using a raspberry Pi and LED

Grant-Jacob, James A. (2025) Deep learning-based pollen imaging using a raspberry Pi and LED. Southampton Imaging Conference | Light on Life, Avenue Campus, Southampton, United Kingdom. (Submitted)

Record type: Conference or Workshop Item (Other)

Abstract

Developing an affordable, compact imaging sensor could significantly enhance global airborne pollen monitoring and alleviate hay fever symptoms. By using a white light LED to illuminate pollen grains and capturing their scattering patterns with a Raspberry Pi camera, we can transform these patterns into detailed images through deep learning techniques. Our method successfully generates images of pollen from plant species not included in the neural network's training data. This technique could also be applied to imaging fungal spores and airborne particulates that contribute to air pollution, offering valuable insights in environmental science, health science, and agriculture. Furthermore, it could help develop more efficient air quality monitoring systems and support research on the effects of airborne particles on human health and crop productivity. By providing detailed images and data on various airborne particulates, this approach can enhance our understanding of how these particles interact with the environment and affect respiratory health, allergies, and diseases. Additionally, it can contribute to agricultural research by examining the impact of pollen and other particulates on crop growth and yield, ultimately aiding in the development of strategies to improve air quality and agricultural productivity.

This record has no associated files available for download.