Communication Dans Un Congrès Année : 2023

Toward dust monitoring using Aerōtape, an IoT image processing -based aerosol analyzer

Vers une surveillance des poussières à l'aide d'Aerōtape, un analyseur d'aérosols communicant basé sur le traitement d'images

Résumé

Atmospheric mineral dusts influence the climate, biogeochemistry and air quality. Thus, concentration monitoring of dust particles is of interest for many stakeholders such as the scientific community, medical doctors or the public health authorities. Most commonly used methods to identify airborne particle origin and chemical nature are usually chemical analysis (ICP-MS) of samples collected on quartz filters. Another widely used technique relies on the coupling between particulate number measurement and weather data (measured and simulated). However, the first method suffers from poor time resolution and is expensive, and thus is not suitable for long-term monitoring, and the second has to be calibrated using the first one. In order to tackle these drawbacks, we investigate the ability of our IoT instrument, Aerotape, to monitor dust concentration in real time. Historically dedicated to bioaerosol monitoring, the Aerotape measurement technique is based on image processing (machine learning) of inertially deposited aerosol particles (1 – 200 μm). An instrument has been deployed at Cyprus Institute, where an aerosol instrument platform is operated to monitor desert dust events. The first phase of this experimental campaign was the comparison between Aerotape measurement with other instrument’s (TEOM and PALAS Fidas), which showed excellent agreements. Then, a collection of particle images has been built in order to train a machine learning algorithm to recognize atmospheric dust in an automated manner. The relevancy of the training data selection is ensured by the monitoring of the local measurement station. This methodology paves the way to desert particle source apportionment using the Aerotape and to use it as a network.

EMME-CARE-workshop-2023_v2.pdf (1.55 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-05104049 , version 1 (09-06-2025)

Identifiants

  • HAL Id : hal-05104049 , version 1

Citer

A Reynaud, P.-Y Quéhé, R Bernard, D Filippi, J Sciare, et al.. Toward dust monitoring using Aerōtape, an IoT image processing -based aerosol analyzer. EMME-CARE 3RD ANNUAL ONLINE WORKSHOP, THE CYPRUS INSTITUTE, Nov 2023, Nicosi- Cyprus, Cyprus. ⟨hal-05104049⟩
80 Consultations
22 Téléchargements

Partager

  • More