Cargando…

Fully Convolutional Neural Network for Rapid Flood Segmentation in Synthetic Aperture Radar Imagery

Rapid response to natural hazards, such as floods, is essential to mitigate loss of life and the reduction of suffering. For emergency response teams, access to timely and accurate data is essential. Satellite imagery offers a rich source of information which can be analysed to help determine region...

Descripción completa

Detalles Bibliográficos
Autores principales: Nemni, Edoardo, Bullock, Joseph, Belabbes, Samir, Bromley, Lars
Lenguaje:eng
Publicado: 2020
Materias:
Acceso en línea:https://dx.doi.org/10.3390/rs12162532
http://cds.cern.ch/record/2799352
_version_ 1780972537921929216
author Nemni, Edoardo
Bullock, Joseph
Belabbes, Samir
Bromley, Lars
author_facet Nemni, Edoardo
Bullock, Joseph
Belabbes, Samir
Bromley, Lars
author_sort Nemni, Edoardo
collection CERN
description Rapid response to natural hazards, such as floods, is essential to mitigate loss of life and the reduction of suffering. For emergency response teams, access to timely and accurate data is essential. Satellite imagery offers a rich source of information which can be analysed to help determine regions affected by a disaster. Much remote sensing flood analysis is semi-automated, with time consuming manual components requiring hours to complete. In this study, we present a fully automated approach to the rapid flood mapping currently carried out by many non-governmental, national and international organisations. We design a Convolutional Neural Network (CNN) based method which isolates the flooded pixels in freely available Copernicus Sentinel-1 Synthetic Aperture Radar (SAR) imagery, requiring no optical bands and minimal pre-processing. We test a variety of CNN architectures and train our models on flood masks generated using a combination of classical semi-automated techniques and extensive manual cleaning and visual inspection. Our methodology reduces the time required to develop a flood map by 80%, while achieving strong performance over a wide range of locations and environmental conditions. Given the open-source data and the minimal image cleaning required, this methodology can also be integrated into end-to-end pipelines for more timely and continuous flood monitoring.
id cern-2799352
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2020
record_format invenio
spelling cern-27993522022-01-25T11:03:45Zdoi:10.3390/rs12162532http://cds.cern.ch/record/2799352engNemni, EdoardoBullock, JosephBelabbes, SamirBromley, LarsFully Convolutional Neural Network for Rapid Flood Segmentation in Synthetic Aperture Radar ImageryEngineeringRapid response to natural hazards, such as floods, is essential to mitigate loss of life and the reduction of suffering. For emergency response teams, access to timely and accurate data is essential. Satellite imagery offers a rich source of information which can be analysed to help determine regions affected by a disaster. Much remote sensing flood analysis is semi-automated, with time consuming manual components requiring hours to complete. In this study, we present a fully automated approach to the rapid flood mapping currently carried out by many non-governmental, national and international organisations. We design a Convolutional Neural Network (CNN) based method which isolates the flooded pixels in freely available Copernicus Sentinel-1 Synthetic Aperture Radar (SAR) imagery, requiring no optical bands and minimal pre-processing. We test a variety of CNN architectures and train our models on flood masks generated using a combination of classical semi-automated techniques and extensive manual cleaning and visual inspection. Our methodology reduces the time required to develop a flood map by 80%, while achieving strong performance over a wide range of locations and environmental conditions. Given the open-source data and the minimal image cleaning required, this methodology can also be integrated into end-to-end pipelines for more timely and continuous flood monitoring.oai:cds.cern.ch:27993522020
spellingShingle Engineering
Nemni, Edoardo
Bullock, Joseph
Belabbes, Samir
Bromley, Lars
Fully Convolutional Neural Network for Rapid Flood Segmentation in Synthetic Aperture Radar Imagery
title Fully Convolutional Neural Network for Rapid Flood Segmentation in Synthetic Aperture Radar Imagery
title_full Fully Convolutional Neural Network for Rapid Flood Segmentation in Synthetic Aperture Radar Imagery
title_fullStr Fully Convolutional Neural Network for Rapid Flood Segmentation in Synthetic Aperture Radar Imagery
title_full_unstemmed Fully Convolutional Neural Network for Rapid Flood Segmentation in Synthetic Aperture Radar Imagery
title_short Fully Convolutional Neural Network for Rapid Flood Segmentation in Synthetic Aperture Radar Imagery
title_sort fully convolutional neural network for rapid flood segmentation in synthetic aperture radar imagery
topic Engineering
url https://dx.doi.org/10.3390/rs12162532
http://cds.cern.ch/record/2799352
work_keys_str_mv AT nemniedoardo fullyconvolutionalneuralnetworkforrapidfloodsegmentationinsyntheticapertureradarimagery
AT bullockjoseph fullyconvolutionalneuralnetworkforrapidfloodsegmentationinsyntheticapertureradarimagery
AT belabbessamir fullyconvolutionalneuralnetworkforrapidfloodsegmentationinsyntheticapertureradarimagery
AT bromleylars fullyconvolutionalneuralnetworkforrapidfloodsegmentationinsyntheticapertureradarimagery