TY - JOUR
T1 - Beyond here and now
T2 - Evaluating pollution estimation across space and time from street view images with deep learning
AU - Nathvani, Ricky
AU - D., Vishwanath
AU - Clark, Sierra N.
AU - Alli, Abosede S.
AU - Muller, Emily
AU - Coste, Henri
AU - Bennett, James E.
AU - Nimo, James
AU - Moses, Josephine Bedford
AU - Baah, Solomon
AU - Hughes, Allison
AU - Suel, Esra
AU - Metzler, Antje Barbara
AU - Rashid, Theo
AU - Brauer, Michael
AU - Baumgartner, Jill
AU - Owusu, George
AU - Agyei-Mensah, Samuel
AU - Arku, Raphael E.
AU - Ezzati, Majid
N1 - Publisher Copyright:
© 2023 The Authors
PY - 2023/12/10
Y1 - 2023/12/10
N2 - Advances in computer vision, driven by deep learning, allows for the inference of environmental pollution and its potential sources from images. The spatial and temporal generalisability of image-based pollution models is crucial in their real-world application, but is currently understudied, particularly in low-income countries where infrastructure for measuring the complex patterns of pollution is limited and modelling may therefore provide the most utility. We employed convolutional neural networks (CNNs) for two complementary classification models, in both an end-to-end approach and as an interpretable feature extractor (object detection), to estimate spatially and temporally resolved fine particulate matter (PM2.5) and noise levels in Accra, Ghana. Data used for training the models were from a unique dataset of over 1.6 million images collected over 15 months at 145 representative locations across the city, paired with air and noise measurements. Both end-to-end CNN and object-based approaches surpassed null model benchmarks for predicting PM2.5 and noise at single locations, but performance deteriorated when applied to other locations. Model accuracy diminished when tested on images from locations unseen during training, but improved by sampling a greater number of locations during model training, even if the total quantity of data was reduced. The end-to-end models used characteristics of images associated with atmospheric visibility for predicting PM2.5, and specific objects such as vehicles and people for noise. The results demonstrate the potential and challenges of image-based, spatiotemporal air pollution and noise estimation, and that robust, environmental modelling with images requires integration with traditional sensor networks.
AB - Advances in computer vision, driven by deep learning, allows for the inference of environmental pollution and its potential sources from images. The spatial and temporal generalisability of image-based pollution models is crucial in their real-world application, but is currently understudied, particularly in low-income countries where infrastructure for measuring the complex patterns of pollution is limited and modelling may therefore provide the most utility. We employed convolutional neural networks (CNNs) for two complementary classification models, in both an end-to-end approach and as an interpretable feature extractor (object detection), to estimate spatially and temporally resolved fine particulate matter (PM2.5) and noise levels in Accra, Ghana. Data used for training the models were from a unique dataset of over 1.6 million images collected over 15 months at 145 representative locations across the city, paired with air and noise measurements. Both end-to-end CNN and object-based approaches surpassed null model benchmarks for predicting PM2.5 and noise at single locations, but performance deteriorated when applied to other locations. Model accuracy diminished when tested on images from locations unseen during training, but improved by sampling a greater number of locations during model training, even if the total quantity of data was reduced. The end-to-end models used characteristics of images associated with atmospheric visibility for predicting PM2.5, and specific objects such as vehicles and people for noise. The results demonstrate the potential and challenges of image-based, spatiotemporal air pollution and noise estimation, and that robust, environmental modelling with images requires integration with traditional sensor networks.
KW - Air pollution
KW - Computer vision
KW - Deep learning
KW - Environmental modelling
KW - Noise pollution
KW - Street-view images
UR - http://www.scopus.com/inward/record.url?scp=85171344363&partnerID=8YFLogxK
U2 - 10.1016/j.scitotenv.2023.166168
DO - 10.1016/j.scitotenv.2023.166168
M3 - Article
C2 - 37586538
AN - SCOPUS:85171344363
SN - 0048-9697
VL - 903
JO - Science of the Total Environment
JF - Science of the Total Environment
M1 - 166168
ER -