Imperial College London
Portrait Picture

DrRickyNathvani

Faculty of Medicine, School of Public Health

Early Career Research Fellow
 
 
 
//

Contact

 

r.nathvani

 
 
//

Location

 

UREN.1118Sir Michael Uren HubWhite City Campus

//

Summary

 

Publications

Citation

BibTex format

@article{Metzler:2023:10.1016/j.scitotenv.2023.164794,
author = {Metzler, AB and Nathvani, R and Sharmanska, V and Bai, W and Muller, E and Moulds, S and Agyei-Asabere, C and Adjei-Boadih, D and Kyere-Gyeabour, E and Tetteh, JD and Owusu, G and Agyei-Mensah, S and Baumgartner, J and Robinson, BE and Arku, RE and Ezzati, M},
doi = {10.1016/j.scitotenv.2023.164794},
journal = {Science of the Total Environment},
pages = {1--14},
title = {Phenotyping urban built and natural environments with high-resolution satellite images and unsupervised deep learning},
url = {http://dx.doi.org/10.1016/j.scitotenv.2023.164794},
volume = {893},
year = {2023}
}
Download

RIS format (EndNote, RefMan)

TY  - JOUR
AB  - Cities in the developing world are expanding rapidly, and undergoing changes to their roads, buildings, vegetation, and other land use characteristics. Timely data are needed to ensure that urban change enhances health, wellbeing and sustainability. We present and evaluate a novel unsupervised deep clustering method to classify and characterise the complex and multidimensional built and natural environments of cities into interpretable clusters using high-resolution satellite images. We applied our approach to a high-resolution (0.3 m/pixel) satellite image of Accra, Ghana, one of the fastest growing cities in sub-Saharan Africa, and contextualised the results with demographic and environmental data that were not used for clustering. We show that clusters obtained solely from images capture distinct interpretable phenotypes of the urban natural (vegetation and water) and built (building count, size, density, and orientation; length and arrangement of roads) environment, and population, either as a unique defining characteristic (e.g., bodies of water or dense vegetation) or in combination (e.g., buildings surrounded by vegetation or sparsely populated areas intermixed with roads). Clusters that were based on a single defining characteristic were robust to the spatial scale of analysis and the choice of cluster number, whereas those based on a combination of characteristics changed based on scale and number of clusters. The results demonstrate that satellite data and unsupervised deep learning provide a cost-effective, interpretable and scalable approach for real-time tracking of sustainable urban development, especially where traditional environmental and demographic data are limited and infrequent.
AU  - Metzler,AB
AU  - Nathvani,R
AU  - Sharmanska,V
AU  - Bai,W
AU  - Muller,E
AU  - Moulds,S
AU  - Agyei-Asabere,C
AU  - Adjei-Boadih,D
AU  - Kyere-Gyeabour,E
AU  - Tetteh,JD
AU  - Owusu,G
AU  - Agyei-Mensah,S
AU  - Baumgartner,J
AU  - Robinson,BE
AU  - Arku,RE
AU  - Ezzati,M
DO  - 10.1016/j.scitotenv.2023.164794
EP  - 14
PY  - 2023///
SN  - 0048-9697
SP  - 1
TI  - Phenotyping urban built and natural environments with high-resolution satellite images and unsupervised deep learning
T2  - Science of the Total Environment
UR  - http://dx.doi.org/10.1016/j.scitotenv.2023.164794
UR  - https://www.sciencedirect.com/science/article/pii/S0048969723034174
UR  - http://hdl.handle.net/10044/1/104875
VL  - 893
ER  -
Download