Automatic detection of airborne pollen: an overview

Abstract

Pollen monitoring has traditionally been carried out using manual methods first developed in the early 1950s. Although this technique has been recently standardised, it suffers from several drawbacks, notably data usually only being available with a delay of 3–9 days and usually delivered at a daily resolution. Several automatic instruments have come on to the market over the past few years, with more new devices also under development. This paper provides a comprehensive overview of all available and developing automatic instruments, how they measure, how they identify airborne pollen, what impacts measurement quality, as well as what potential there is for further advancement in the field of bioaerosol monitoring.

A Modified Spectroscopic Approach for the Real-Time Detection of Pollen and Fungal Spores at a Semi-Urban Site Using the WIBS-4+, Part I

“…Such deviations could be elucidated due to differences in instrument operating principles. The Hirst in particular has been used extensively since the 1950s, despite its historical popularity, there are a number of drawbacks in using the Hirst that can lead to deviances when compared to real-time devices [77]. For one, the PBAP concentrations determined by the Hirst method are based on the statistical extrapolation of a portion of the counted sample slide.…”

“…As previously mentioned, the enhanced flow rate of the Hirst makes it more efficient at sampling larger (potentially faster moving) particles that are unaffected and thus under sampled by the lower flow rate of the WIBS. Meteorological conditions such as strong wind speeds could also influence the sampling efficiency of the instruments [77,79]. By design the Hirst sampler continually orientates itself to face the prevailing wind.…”

Explainable AI for unveiling deep learning pollen classification model based on fusion of scattered light patterns and fluorescence spectroscopy

“…Several automatic instruments for pollen classification emerged on the market 25 that are based on digital images or electrical signals from various types of sensors. Such instruments further demand machine learning algorithms to automate the process of pollen classification.…”

“…Degree of shrivelling, and resulting size change, could affect performance of the model utilizing scattering signal. It is known that pollen dehydration and subsequent rehydration occurs within minutes 25 and the character of this process is unknown. Specific hydration level pollen www.nature.com/scientificreports/ could be more easily classified using scattering signal but to evaluate such hypothesis controlled conditions in the laboratory would be required.…”

Automatic real-time monitoring of fungal spores: the case of Alternaria spp.

“…The automation of operational pollen monitoring networks has been taking place over the past few years (Buters et al., 2022 ; Crouzy et al, 2016 ; Tešendić et al, 2022 ). Recent instruments deliver airborne pollen concentrations in real-time at high-temporal resolutions, better corresponding to the active dynamics of meteorology that influence particle dispersion.…”