Data enrichment and calibration for PM 2.5 low-cost optical sensors
Апстракт
Particulate matter (PM) in air has been proven to be hazardous to human health. Until recently, monitoring of air quality has been done by professional agencies. Nowadays, the availability of portable, low cost microsensor devices and the exponential growth of IoT (Internet of Things) in everyday life has enabled widespread monitoring of air quality among all citizens.[1]. For PM measurements, optical sensors measure light scattering by particles carried in an air stream through a light beam, which is converted by computation to equivalent mass concentration. Light scattering is strongly affected by parameters such as particle density, particle hygroscopicity, refraction index, and particle composition [2]. In this study, we measured PM 2.5 by seven AQ MESH low-cost optical sensors and compared the measured data with the ones obtained from the reference monitoring station (SEPA). Could we, by a sequence of low-processing data enrichment and a simple calibration method, reach an accurac...y as close as a calibration based on machine learning? To answer this question, we used low-processing data enrichment such as resampling, encoding periodic timerelated features and making a composition of the initial low-cost signal at different time scales. We compared two algorithms for the calibration: multivariate linear regression and random forest. The results gave promising results and encouraged us in researching further about signal low-processing to achieve the required quality of data from low-cost sensor devices monitoring air quality [3].
Извор:
16th Photonics Workshop : Book of abstracts, 2023, 40-40Издавач:
- Belgrade : Institute of Physics
Финансирање / пројекти:
- Министарство науке, технолошког развоја и иновација Републике Србије, институционално финансирање - 200017 (Универзитет у Београду, Институт за нуклеарне науке Винча, Београд-Винча) (RS-MESTD-inst-2020-200017)
- EU H2020 Framework Programme [Grant agreement no. 952433 (VIDIS)]
- EU 7th Framework Programme for research, technological development and demonstration [Grant agreement no 308524 (CITI-SENSE)]
Напомена:
- XVI Photonics Workshop : Book of abstracts; March 12-15, 2023; Kopaonik, Serbia
Колекције
Институција/група
VinčaTY - CONF AU - Stojanović, Danka B. AU - Kleut, Duška AU - Davidović, Miloš AU - Lepioufle, Jean-Marie PY - 2023 UR - https://vinar.vin.bg.ac.rs/handle/123456789/13047 AB - Particulate matter (PM) in air has been proven to be hazardous to human health. Until recently, monitoring of air quality has been done by professional agencies. Nowadays, the availability of portable, low cost microsensor devices and the exponential growth of IoT (Internet of Things) in everyday life has enabled widespread monitoring of air quality among all citizens.[1]. For PM measurements, optical sensors measure light scattering by particles carried in an air stream through a light beam, which is converted by computation to equivalent mass concentration. Light scattering is strongly affected by parameters such as particle density, particle hygroscopicity, refraction index, and particle composition [2]. In this study, we measured PM 2.5 by seven AQ MESH low-cost optical sensors and compared the measured data with the ones obtained from the reference monitoring station (SEPA). Could we, by a sequence of low-processing data enrichment and a simple calibration method, reach an accuracy as close as a calibration based on machine learning? To answer this question, we used low-processing data enrichment such as resampling, encoding periodic timerelated features and making a composition of the initial low-cost signal at different time scales. We compared two algorithms for the calibration: multivariate linear regression and random forest. The results gave promising results and encouraged us in researching further about signal low-processing to achieve the required quality of data from low-cost sensor devices monitoring air quality [3]. PB - Belgrade : Institute of Physics C3 - 16th Photonics Workshop : Book of abstracts T1 - Data enrichment and calibration for PM 2.5 low-cost optical sensors SP - 40 EP - 40 UR - https://hdl.handle.net/21.15107/rcub_vinar_13047 ER -
@conference{ author = "Stojanović, Danka B. and Kleut, Duška and Davidović, Miloš and Lepioufle, Jean-Marie", year = "2023", abstract = "Particulate matter (PM) in air has been proven to be hazardous to human health. Until recently, monitoring of air quality has been done by professional agencies. Nowadays, the availability of portable, low cost microsensor devices and the exponential growth of IoT (Internet of Things) in everyday life has enabled widespread monitoring of air quality among all citizens.[1]. For PM measurements, optical sensors measure light scattering by particles carried in an air stream through a light beam, which is converted by computation to equivalent mass concentration. Light scattering is strongly affected by parameters such as particle density, particle hygroscopicity, refraction index, and particle composition [2]. In this study, we measured PM 2.5 by seven AQ MESH low-cost optical sensors and compared the measured data with the ones obtained from the reference monitoring station (SEPA). Could we, by a sequence of low-processing data enrichment and a simple calibration method, reach an accuracy as close as a calibration based on machine learning? To answer this question, we used low-processing data enrichment such as resampling, encoding periodic timerelated features and making a composition of the initial low-cost signal at different time scales. We compared two algorithms for the calibration: multivariate linear regression and random forest. The results gave promising results and encouraged us in researching further about signal low-processing to achieve the required quality of data from low-cost sensor devices monitoring air quality [3].", publisher = "Belgrade : Institute of Physics", journal = "16th Photonics Workshop : Book of abstracts", title = "Data enrichment and calibration for PM 2.5 low-cost optical sensors", pages = "40-40", url = "https://hdl.handle.net/21.15107/rcub_vinar_13047" }
Stojanović, D. B., Kleut, D., Davidović, M.,& Lepioufle, J.. (2023). Data enrichment and calibration for PM 2.5 low-cost optical sensors. in 16th Photonics Workshop : Book of abstracts Belgrade : Institute of Physics., 40-40. https://hdl.handle.net/21.15107/rcub_vinar_13047
Stojanović DB, Kleut D, Davidović M, Lepioufle J. Data enrichment and calibration for PM 2.5 low-cost optical sensors. in 16th Photonics Workshop : Book of abstracts. 2023;:40-40. https://hdl.handle.net/21.15107/rcub_vinar_13047 .
Stojanović, Danka B., Kleut, Duška, Davidović, Miloš, Lepioufle, Jean-Marie, "Data enrichment and calibration for PM 2.5 low-cost optical sensors" in 16th Photonics Workshop : Book of abstracts (2023):40-40, https://hdl.handle.net/21.15107/rcub_vinar_13047 .