Imaging pollen using a Raspberry Pi and LED with deep learning.

Journal: The Science of the total environment
PMID:

Abstract

The production of low-cost, small footprint imaging sensor would be invaluable for airborne global monitoring of pollen, which could allow for mitigation of hay fever symptoms. We demonstrate the use of a white light LED (light emitting diode) to illuminate pollen grains and capture their scattering pattern using a Raspberry Pi camera. The scattering patterns are transformed into 20× microscope magnification equivalent images using deep learning. We show the ability to produce images of pollen from plant species previously unseen by the neural network in training. Such a technique could be applied to imaging airborne particulates that contribute to air pollution, and could be used in the field of environmental science, health science and agriculture.

Authors

  • Ben Mills
    Optoelectronics Research Centre, Faculty of Engineering and Physical Sciences, University of Southampton, Southampton, SO17 1BJ, United Kingdom.
  • Michalis N Zervas
    Optoelectronics Research Centre, University of Southampton, Southampton, SO17 1BJ, UK.
  • James A Grant-Jacob
    Optoelectronics Research Centre, Faculty of Engineering and Physical Sciences, University of Southampton, Southampton, SO17 1BJ, United Kingdom.