arrow
Return

Decoding (urban) form and function using spatially explicit deep learning

delete2024-09-01
delete0
delete
OA
AI
M
Martin Fleischmann *
D
Daniel Arribas‐Bel
DOI:10.1016/j.compenvurbsys.2024.102147delete
deleteOriginal
deleteShare
deleteSave
View PDF
Abstract

Abstract

En 中文
This paper explores how can geographical dimension be incorporated into deep learning designed to understand the composition of urban landscapes based on Sentinel 2 satellite imagery. Compared to standard computer vision, satellite imagery is unique as images sampled from the data form a continuous array, rather than being fully independent. We argue that the spatial configuration of the images is as important as the content of each image when attempting to capture a pattern that reflects the structure of the urban environment. We propose a series of approaches explicitly incorporating spatial dimension in the predictive pipeline based on the EfficientNetB4 convolutional neural network (CNN) and experimentally test their effect on model performance. The experiments in this study cover the scale of the sampled area, the effect of spatial augmentation, and the role of modelling (logit ensemble and histogram-based gradient-boosted classifiers) with and without the spatial context on the outputs of the neural network-generated vector of probabilities while trying to predict spatial signatures, a classification of primarily urban landscape based on form and function. The results suggest that certain ways of embedding spatial information, especially in the modelling step, consistently significantly improve the prediction accuracy and shall be considered on top of standard CNNs.
Keywords:
Spatial signatures
Classification
Remote sensing
Artificial intelligence
Open data

Journal

Computers Environment and Urban Systems cover
Computers Environment and Urban Systems
IF:
8.3
Papers:
1.6K
Citations:
1.1W

Organization

U
University of Liverpool
Scholars:
1.8K
Papers: 758
Citations: 1.9W
C
Charles University Prague
Scholars:
3.0W
Papers: 2.2W
Citations: 2.4W
Cited Papers

Cited Papers

BODY HISTORY
err1986-12-01
err0
PREAI
errIvan Illich
errShare
errSave
On the knowledge gain of urban morphology from space
err2022-07-01
err15
errOAAI
errWang, Jiong; Georganos, Stefanos; Kuffer, Monika; Abascal, Angela; Vanhuysse, Sabine
errShare
errSave
Lactose in Animal and Human Feeding: A Review
err1957-09-01
err0
errOAAI
errR.L. Atkinson; F.H. Kratzer; G.F. Stewart
errShare
errSave
Across-Line SNP Association Study for Direct and Associative Effects on Feather Damage in Laying Hens
err2010-05-23
err0
PREAI
errF. Biscarini; H. Bovenhuis; J. van der Poel; T. B. Rodenburg; A. P. Jungerius; J. A. M. van Arendonk
errShare
errSave
Mapping the local climate zones of urban areas by GIS-based and WUDAPT methods: A case study of Hong Kong
err2018-06-01
err151
PREAI
errWang, Ran; Ren, Chao; Xu, Yong; Lau, Kevin Ka-Lun; Shi, Yuan
errShare
errSave
Understanding (urban) spaces through form and function
err2022-10-01
err27
errOAAI
errArribas-Bel, Daniel; Fleischmann, Martin
errShare
errSave
Seven city types representing morphologic configurations of cities across the globe
errCITIES
IF6.6
err2020-10-01
err60
errOAAI
errTaubenboeck, H.; Debray, H.; Qiu, C.; Schmitt, M.; Wang, Y.; Zhu, X. X.
errShare
errSave
researcher View more