Indoor Air Pollution Source Localization Based on Small-Sample Training Convolutional Neural Networks

In addressing the problem of indoor air pollution source localization, traditional methods have limitations such as strong sample dependence and low computational efficiency. This study uses a convolutional neural network to establish a pollution source inversion method based on small samples. By in...

Full description

Saved in:
Bibliographic Details
Main Authors: Tiancheng Ye, Mengtao Han
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Buildings
Subjects:
Online Access:https://www.mdpi.com/2075-5309/15/8/1244
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849712504101404672
author Tiancheng Ye
Mengtao Han
author_facet Tiancheng Ye
Mengtao Han
author_sort Tiancheng Ye
collection DOAJ
description In addressing the problem of indoor air pollution source localization, traditional methods have limitations such as strong sample dependence and low computational efficiency. This study uses a convolutional neural network to establish a pollution source inversion method based on small samples. By integrating computational fluid dynamics simulation data and deep learning techniques, a spatial pollution source identification model suitable for limited-sample conditions was constructed. In a benchmark scenario, the optimized model achieved a localization of 82.3% weighted accuracy within a prediction radius of 1 m, and the corresponding normalized error of the detected area was of less than 0.26%. In cross-scenario verification, the localization accuracy within a 1 m radius increased to 100%, and the corresponding predicted Euclidean distance error decreased by 21.43%. By using the optimal cutting ratio (α = 0.25) and a rotation-enhanced dataset (θ = 10°, n = 36), the model reduced the cross-space sample requirement to 1/5 of that of the benchmark scenario while ensuring the accuracy of spatial representation. The research findings provide an efficient and reliable deep learning solution for the localization of pollution sources in complex spaces.
format Article
id doaj-art-7246afff6950426b954ac52cc4580366
institution DOAJ
issn 2075-5309
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Buildings
spelling doaj-art-7246afff6950426b954ac52cc45803662025-08-20T03:14:15ZengMDPI AGBuildings2075-53092025-04-01158124410.3390/buildings15081244Indoor Air Pollution Source Localization Based on Small-Sample Training Convolutional Neural NetworksTiancheng Ye0Mengtao Han1School of Architecture and Urban Planning, Huazhong University of Science and Technology, Wuhan 430074, ChinaSchool of Architecture and Urban Planning, Huazhong University of Science and Technology, Wuhan 430074, ChinaIn addressing the problem of indoor air pollution source localization, traditional methods have limitations such as strong sample dependence and low computational efficiency. This study uses a convolutional neural network to establish a pollution source inversion method based on small samples. By integrating computational fluid dynamics simulation data and deep learning techniques, a spatial pollution source identification model suitable for limited-sample conditions was constructed. In a benchmark scenario, the optimized model achieved a localization of 82.3% weighted accuracy within a prediction radius of 1 m, and the corresponding normalized error of the detected area was of less than 0.26%. In cross-scenario verification, the localization accuracy within a 1 m radius increased to 100%, and the corresponding predicted Euclidean distance error decreased by 21.43%. By using the optimal cutting ratio (α = 0.25) and a rotation-enhanced dataset (θ = 10°, n = 36), the model reduced the cross-space sample requirement to 1/5 of that of the benchmark scenario while ensuring the accuracy of spatial representation. The research findings provide an efficient and reliable deep learning solution for the localization of pollution sources in complex spaces.https://www.mdpi.com/2075-5309/15/8/1244pollution source inversionconvolutional neural network (CNN)deep learningsmall samplespatial localization accuracy
spellingShingle Tiancheng Ye
Mengtao Han
Indoor Air Pollution Source Localization Based on Small-Sample Training Convolutional Neural Networks
Buildings
pollution source inversion
convolutional neural network (CNN)
deep learning
small sample
spatial localization accuracy
title Indoor Air Pollution Source Localization Based on Small-Sample Training Convolutional Neural Networks
title_full Indoor Air Pollution Source Localization Based on Small-Sample Training Convolutional Neural Networks
title_fullStr Indoor Air Pollution Source Localization Based on Small-Sample Training Convolutional Neural Networks
title_full_unstemmed Indoor Air Pollution Source Localization Based on Small-Sample Training Convolutional Neural Networks
title_short Indoor Air Pollution Source Localization Based on Small-Sample Training Convolutional Neural Networks
title_sort indoor air pollution source localization based on small sample training convolutional neural networks
topic pollution source inversion
convolutional neural network (CNN)
deep learning
small sample
spatial localization accuracy
url https://www.mdpi.com/2075-5309/15/8/1244
work_keys_str_mv AT tianchengye indoorairpollutionsourcelocalizationbasedonsmallsampletrainingconvolutionalneuralnetworks
AT mengtaohan indoorairpollutionsourcelocalizationbasedonsmallsampletrainingconvolutionalneuralnetworks