Spatial Modeling of Airborne Particles (PM2.5 and PM10) in Tehran city Using Convolutional Neural Network.

Air pollution poses significant risks to human health and the environment, which makes it necessary to create effective strategies for air quality management. This study presents an approach for air quality management in Tehran using the Convolutional Neural Network (CNN) algorithm. The proposed met...

Full description

Saved in:
Bibliographic Details
Main Authors: Abed Bashar Doost, Mohammad Saadi Mesgari
Format: Article
Language:fas
Published: I.R. of Iran Meteorological Organization 2024-03-01
Series:Nīvār
Subjects:
Online Access:https://nivar.irimo.ir/article_192602_7ef07a445c311907bf6f9e2ab32a61a1.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1841559943840268288
author Abed Bashar Doost
Mohammad Saadi Mesgari
author_facet Abed Bashar Doost
Mohammad Saadi Mesgari
author_sort Abed Bashar Doost
collection DOAJ
description Air pollution poses significant risks to human health and the environment, which makes it necessary to create effective strategies for air quality management. This study presents an approach for air quality management in Tehran using the Convolutional Neural Network (CNN) algorithm. The proposed method provides the possibility of spatial modeling and preparation of risk maps of two important air pollutants, namely particulate matter 2.5 (PM2.5) and particulate matter 10 (PM10). To develop this air pollution model, the data available in the database containing the annual average of two pollutants from 2012 to 2022 were used. In this model, various parameters affecting air pollution including altitude, humidity, distance to industrial areas, normalized difference index of plants (NDVI), population density, precipitation, distance to the street, temperature, traffic volume, wind direction, and wind speed are considered. Taken and spatial modeling of two pollutants using CNN has been done. The evaluation of the model was done using different evaluation criteria, and the findings showed that the R-squared (R2) values in this model for PM2.5 and PM10 pollutants are 0.889 and 0.972, respectively. The accuracy of the risk map was evaluated using relative operating characteristic (ROC) for two pollutants, and the findings showed that the CNN model has an acceptable accuracy in producing the pollution risk map. In general, risk maps provide useful information about geographic areas with high pollution risks and help in decision-making and targeted pollution reduction efforts.
format Article
id doaj-art-594578d6304843968ed12a8855aea50f
institution Kabale University
issn 1735-0565
2645-3347
language fas
publishDate 2024-03-01
publisher I.R. of Iran Meteorological Organization
record_format Article
series Nīvār
spelling doaj-art-594578d6304843968ed12a8855aea50f2025-01-05T10:55:06ZfasI.R. of Iran Meteorological OrganizationNīvār1735-05652645-33472024-03-0148124-125314910.30467/nivar.2024.430255.1276192602Spatial Modeling of Airborne Particles (PM2.5 and PM10) in Tehran city Using Convolutional Neural Network.Abed Bashar Doost0Mohammad Saadi Mesgari1Doctoral student of civil engineering and mapping, geographic information systems, K. N. Toosi University, Tehran, Iran.Associate Professor of Department of Geodesy and Geomatics, K. N. Toosi University of Technology, Tehran, IranAir pollution poses significant risks to human health and the environment, which makes it necessary to create effective strategies for air quality management. This study presents an approach for air quality management in Tehran using the Convolutional Neural Network (CNN) algorithm. The proposed method provides the possibility of spatial modeling and preparation of risk maps of two important air pollutants, namely particulate matter 2.5 (PM2.5) and particulate matter 10 (PM10). To develop this air pollution model, the data available in the database containing the annual average of two pollutants from 2012 to 2022 were used. In this model, various parameters affecting air pollution including altitude, humidity, distance to industrial areas, normalized difference index of plants (NDVI), population density, precipitation, distance to the street, temperature, traffic volume, wind direction, and wind speed are considered. Taken and spatial modeling of two pollutants using CNN has been done. The evaluation of the model was done using different evaluation criteria, and the findings showed that the R-squared (R2) values in this model for PM2.5 and PM10 pollutants are 0.889 and 0.972, respectively. The accuracy of the risk map was evaluated using relative operating characteristic (ROC) for two pollutants, and the findings showed that the CNN model has an acceptable accuracy in producing the pollution risk map. In general, risk maps provide useful information about geographic areas with high pollution risks and help in decision-making and targeted pollution reduction efforts.https://nivar.irimo.ir/article_192602_7ef07a445c311907bf6f9e2ab32a61a1.pdfair pollutantsspatial modelingrisk mapdeep learning
spellingShingle Abed Bashar Doost
Mohammad Saadi Mesgari
Spatial Modeling of Airborne Particles (PM2.5 and PM10) in Tehran city Using Convolutional Neural Network.
Nīvār
air pollutants
spatial modeling
risk map
deep learning
title Spatial Modeling of Airborne Particles (PM2.5 and PM10) in Tehran city Using Convolutional Neural Network.
title_full Spatial Modeling of Airborne Particles (PM2.5 and PM10) in Tehran city Using Convolutional Neural Network.
title_fullStr Spatial Modeling of Airborne Particles (PM2.5 and PM10) in Tehran city Using Convolutional Neural Network.
title_full_unstemmed Spatial Modeling of Airborne Particles (PM2.5 and PM10) in Tehran city Using Convolutional Neural Network.
title_short Spatial Modeling of Airborne Particles (PM2.5 and PM10) in Tehran city Using Convolutional Neural Network.
title_sort spatial modeling of airborne particles pm2 5 and pm10 in tehran city using convolutional neural network
topic air pollutants
spatial modeling
risk map
deep learning
url https://nivar.irimo.ir/article_192602_7ef07a445c311907bf6f9e2ab32a61a1.pdf
work_keys_str_mv AT abedbashardoost spatialmodelingofairborneparticlespm25andpm10intehrancityusingconvolutionalneuralnetwork
AT mohammadsaadimesgari spatialmodelingofairborneparticlespm25andpm10intehrancityusingconvolutionalneuralnetwork