Loading…

Deep-water oil-spill monitoring and recurrence analysis in the Brazilian territory using Sentinel-1 time series and deep learning

•This is the first oil-spill deep learning study in the Brazilian territory.•We compared three deep learning architectures and four backbones.•We detected hotspots using recurrence analysis from deep learning predictions.•A recurrence threshold greater than two eliminates nearly all look-alike featu...

Full description

Saved in:
Bibliographic Details
Published in:International journal of applied earth observation and geoinformation 2022-03, Vol.107, p.102695, Article 102695
Main Authors: de Moura, Nájla Vilar Aires, de Carvalho, Osmar Luiz Ferreira, Gomes, Roberto Arnaldo Trancoso, Guimarães, Renato Fontes, de Carvalho Júnior, Osmar Abílio
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by cdi_FETCH-LOGICAL-c406t-f5861ff1e03164c5825075bb4beaac62df5f3e75fda1fac08a0bb31fdfe505f33
cites cdi_FETCH-LOGICAL-c406t-f5861ff1e03164c5825075bb4beaac62df5f3e75fda1fac08a0bb31fdfe505f33
container_end_page
container_issue
container_start_page 102695
container_title International journal of applied earth observation and geoinformation
container_volume 107
creator de Moura, Nájla Vilar Aires
de Carvalho, Osmar Luiz Ferreira
Gomes, Roberto Arnaldo Trancoso
Guimarães, Renato Fontes
de Carvalho Júnior, Osmar Abílio
description •This is the first oil-spill deep learning study in the Brazilian territory.•We compared three deep learning architectures and four backbones.•We detected hotspots using recurrence analysis from deep learning predictions.•A recurrence threshold greater than two eliminates nearly all look-alike features.•Most oil-spills events in the Campos Basin are related to oil platforms. Oil spills are a worldwide concern since they cause environmental problems and financial losses. Automatic detection plays a crucial role in rapid decision-making to reduce damage. In this context, deep learning and remote sensing are powerful tools with successful applications in many regions. However, there are still no studies on deep-water zones and in the Brazilian territory. The present research has three contributions: (1) create an oil spill dataset on the Brazilian region, (2) compare state-of-the-art deep learning models for this task, and (3) propose a novel application on the Campos Basin (Brazil's largest producer) using Sentinel-1 time series to identify more susceptible regions, encompassing 138 images from 2016 to 2021. The experts manually labeled the dataset with confirmation from the oil company and federal inspection, resulting in 800 images with 512 × 512 spatial dimensions and their respective annotations (600 for training, 125 for validation, and 75 for testing). The study compared three semantic segmentation architectures (U-net, DeepLabv3+, and LinkNet) with four backbones (ResNet-101, ResNet-50, Efficient-net-B0, and Efficient-net-B3) resulting in 12 models. The U-net with the Efficient-net-B3 backbone presented slightly better results (98% accuracy, 75% precision, 78% recall, 76% F-score, and 62% IoU). The time-series analysis considered images with 8,192 × 11,264 spatial dimensions. The entire image classification used a sliding window approach with a 128-pixel stride (averaging the overlapping pixels). Moreover, recurrence analysis detected the more susceptible areas and eliminated the look-alike. False-positive features have a high spatial presence, but with low recurrence (rarely reaching 3 events), which facilitates their elimination using a threshold recurrence. This novel analysis for oil spills is suitable for understanding the patterns along the time and providing regions for the authorities in real-time.
doi_str_mv 10.1016/j.jag.2022.102695
format article
fullrecord <record><control><sourceid>elsevier_doaj_</sourceid><recordid>TN_cdi_doaj_primary_oai_doaj_org_article_7df9b17c0fc14137bceebb3eb51adeb9</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><els_id>S0303243422000216</els_id><doaj_id>oai_doaj_org_article_7df9b17c0fc14137bceebb3eb51adeb9</doaj_id><sourcerecordid>S0303243422000216</sourcerecordid><originalsourceid>FETCH-LOGICAL-c406t-f5861ff1e03164c5825075bb4beaac62df5f3e75fda1fac08a0bb31fdfe505f33</originalsourceid><addsrcrecordid>eNp9kcFO3DAQhiNUJCjwANz8AtnaTuxkxamF0iIhcWgrcbPG9niZyJus7EC1vfHmddiqR072b_v_PDN_VV0KvhJc6E_DaoDNSnIpi5Z6rY6qU9F3su6lfvxQ9kqv675t5En1MeeBc9F1uj-tXm8Qd_VvmDGxiWKddxQj204jzVOiccNg9Cyhe04JR4dFQtxnyoxGNj8h-5LgD0WCojClxbRnz3kx_sBxphFjLdhMW2QZE2F-4_nyJ4sIaSwPz6vjADHjxb_1rPp1-_Xn9ff6_uHb3fXn-9q1XM91UL0WIQjkjdCtU71UvFPWthYBnJY-qNBgp4IHEcDxHri1jQg-oOLlqjmr7g5cP8Fgdom2kPZmAjJvB1PaGEgzuYim82FtRed4cKIVTWcdYoGhVQI82nVhiQPLpSnnhOE_T3Cz5GEGU_IwSx7mkEfxXB08WJp8IUwmO1pm6qnMdy5V0Dvuv1GMl0o</addsrcrecordid><sourcetype>Open Website</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Deep-water oil-spill monitoring and recurrence analysis in the Brazilian territory using Sentinel-1 time series and deep learning</title><source>Elsevier:Jisc Collections:Elsevier Read and Publish Agreement 2022-2024:Freedom Collection (Reading list)</source><creator>de Moura, Nájla Vilar Aires ; de Carvalho, Osmar Luiz Ferreira ; Gomes, Roberto Arnaldo Trancoso ; Guimarães, Renato Fontes ; de Carvalho Júnior, Osmar Abílio</creator><creatorcontrib>de Moura, Nájla Vilar Aires ; de Carvalho, Osmar Luiz Ferreira ; Gomes, Roberto Arnaldo Trancoso ; Guimarães, Renato Fontes ; de Carvalho Júnior, Osmar Abílio</creatorcontrib><description>•This is the first oil-spill deep learning study in the Brazilian territory.•We compared three deep learning architectures and four backbones.•We detected hotspots using recurrence analysis from deep learning predictions.•A recurrence threshold greater than two eliminates nearly all look-alike features.•Most oil-spills events in the Campos Basin are related to oil platforms. Oil spills are a worldwide concern since they cause environmental problems and financial losses. Automatic detection plays a crucial role in rapid decision-making to reduce damage. In this context, deep learning and remote sensing are powerful tools with successful applications in many regions. However, there are still no studies on deep-water zones and in the Brazilian territory. The present research has three contributions: (1) create an oil spill dataset on the Brazilian region, (2) compare state-of-the-art deep learning models for this task, and (3) propose a novel application on the Campos Basin (Brazil's largest producer) using Sentinel-1 time series to identify more susceptible regions, encompassing 138 images from 2016 to 2021. The experts manually labeled the dataset with confirmation from the oil company and federal inspection, resulting in 800 images with 512 × 512 spatial dimensions and their respective annotations (600 for training, 125 for validation, and 75 for testing). The study compared three semantic segmentation architectures (U-net, DeepLabv3+, and LinkNet) with four backbones (ResNet-101, ResNet-50, Efficient-net-B0, and Efficient-net-B3) resulting in 12 models. The U-net with the Efficient-net-B3 backbone presented slightly better results (98% accuracy, 75% precision, 78% recall, 76% F-score, and 62% IoU). The time-series analysis considered images with 8,192 × 11,264 spatial dimensions. The entire image classification used a sliding window approach with a 128-pixel stride (averaging the overlapping pixels). Moreover, recurrence analysis detected the more susceptible areas and eliminated the look-alike. False-positive features have a high spatial presence, but with low recurrence (rarely reaching 3 events), which facilitates their elimination using a threshold recurrence. This novel analysis for oil spills is suitable for understanding the patterns along the time and providing regions for the authorities in real-time.</description><identifier>ISSN: 1569-8432</identifier><identifier>EISSN: 1872-826X</identifier><identifier>DOI: 10.1016/j.jag.2022.102695</identifier><language>eng</language><publisher>Elsevier B.V</publisher><subject>Campos Basin ; Coastal management ; Semantic segmentation</subject><ispartof>International journal of applied earth observation and geoinformation, 2022-03, Vol.107, p.102695, Article 102695</ispartof><rights>2022</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c406t-f5861ff1e03164c5825075bb4beaac62df5f3e75fda1fac08a0bb31fdfe505f33</citedby><cites>FETCH-LOGICAL-c406t-f5861ff1e03164c5825075bb4beaac62df5f3e75fda1fac08a0bb31fdfe505f33</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,780,784,27922,27923</link.rule.ids></links><search><creatorcontrib>de Moura, Nájla Vilar Aires</creatorcontrib><creatorcontrib>de Carvalho, Osmar Luiz Ferreira</creatorcontrib><creatorcontrib>Gomes, Roberto Arnaldo Trancoso</creatorcontrib><creatorcontrib>Guimarães, Renato Fontes</creatorcontrib><creatorcontrib>de Carvalho Júnior, Osmar Abílio</creatorcontrib><title>Deep-water oil-spill monitoring and recurrence analysis in the Brazilian territory using Sentinel-1 time series and deep learning</title><title>International journal of applied earth observation and geoinformation</title><description>•This is the first oil-spill deep learning study in the Brazilian territory.•We compared three deep learning architectures and four backbones.•We detected hotspots using recurrence analysis from deep learning predictions.•A recurrence threshold greater than two eliminates nearly all look-alike features.•Most oil-spills events in the Campos Basin are related to oil platforms. Oil spills are a worldwide concern since they cause environmental problems and financial losses. Automatic detection plays a crucial role in rapid decision-making to reduce damage. In this context, deep learning and remote sensing are powerful tools with successful applications in many regions. However, there are still no studies on deep-water zones and in the Brazilian territory. The present research has three contributions: (1) create an oil spill dataset on the Brazilian region, (2) compare state-of-the-art deep learning models for this task, and (3) propose a novel application on the Campos Basin (Brazil's largest producer) using Sentinel-1 time series to identify more susceptible regions, encompassing 138 images from 2016 to 2021. The experts manually labeled the dataset with confirmation from the oil company and federal inspection, resulting in 800 images with 512 × 512 spatial dimensions and their respective annotations (600 for training, 125 for validation, and 75 for testing). The study compared three semantic segmentation architectures (U-net, DeepLabv3+, and LinkNet) with four backbones (ResNet-101, ResNet-50, Efficient-net-B0, and Efficient-net-B3) resulting in 12 models. The U-net with the Efficient-net-B3 backbone presented slightly better results (98% accuracy, 75% precision, 78% recall, 76% F-score, and 62% IoU). The time-series analysis considered images with 8,192 × 11,264 spatial dimensions. The entire image classification used a sliding window approach with a 128-pixel stride (averaging the overlapping pixels). Moreover, recurrence analysis detected the more susceptible areas and eliminated the look-alike. False-positive features have a high spatial presence, but with low recurrence (rarely reaching 3 events), which facilitates their elimination using a threshold recurrence. This novel analysis for oil spills is suitable for understanding the patterns along the time and providing regions for the authorities in real-time.</description><subject>Campos Basin</subject><subject>Coastal management</subject><subject>Semantic segmentation</subject><issn>1569-8432</issn><issn>1872-826X</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>DOA</sourceid><recordid>eNp9kcFO3DAQhiNUJCjwANz8AtnaTuxkxamF0iIhcWgrcbPG9niZyJus7EC1vfHmddiqR072b_v_PDN_VV0KvhJc6E_DaoDNSnIpi5Z6rY6qU9F3su6lfvxQ9kqv675t5En1MeeBc9F1uj-tXm8Qd_VvmDGxiWKddxQj204jzVOiccNg9Cyhe04JR4dFQtxnyoxGNj8h-5LgD0WCojClxbRnz3kx_sBxphFjLdhMW2QZE2F-4_nyJ4sIaSwPz6vjADHjxb_1rPp1-_Xn9ff6_uHb3fXn-9q1XM91UL0WIQjkjdCtU71UvFPWthYBnJY-qNBgp4IHEcDxHri1jQg-oOLlqjmr7g5cP8Fgdom2kPZmAjJvB1PaGEgzuYim82FtRed4cKIVTWcdYoGhVQI82nVhiQPLpSnnhOE_T3Cz5GEGU_IwSx7mkEfxXB08WJp8IUwmO1pm6qnMdy5V0Dvuv1GMl0o</recordid><startdate>202203</startdate><enddate>202203</enddate><creator>de Moura, Nájla Vilar Aires</creator><creator>de Carvalho, Osmar Luiz Ferreira</creator><creator>Gomes, Roberto Arnaldo Trancoso</creator><creator>Guimarães, Renato Fontes</creator><creator>de Carvalho Júnior, Osmar Abílio</creator><general>Elsevier B.V</general><general>Elsevier</general><scope>6I.</scope><scope>AAFTH</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>DOA</scope></search><sort><creationdate>202203</creationdate><title>Deep-water oil-spill monitoring and recurrence analysis in the Brazilian territory using Sentinel-1 time series and deep learning</title><author>de Moura, Nájla Vilar Aires ; de Carvalho, Osmar Luiz Ferreira ; Gomes, Roberto Arnaldo Trancoso ; Guimarães, Renato Fontes ; de Carvalho Júnior, Osmar Abílio</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c406t-f5861ff1e03164c5825075bb4beaac62df5f3e75fda1fac08a0bb31fdfe505f33</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Campos Basin</topic><topic>Coastal management</topic><topic>Semantic segmentation</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>de Moura, Nájla Vilar Aires</creatorcontrib><creatorcontrib>de Carvalho, Osmar Luiz Ferreira</creatorcontrib><creatorcontrib>Gomes, Roberto Arnaldo Trancoso</creatorcontrib><creatorcontrib>Guimarães, Renato Fontes</creatorcontrib><creatorcontrib>de Carvalho Júnior, Osmar Abílio</creatorcontrib><collection>ScienceDirect Open Access Titles</collection><collection>Elsevier:ScienceDirect:Open Access</collection><collection>CrossRef</collection><collection>DOAJ Directory of Open Access Journals</collection><jtitle>International journal of applied earth observation and geoinformation</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>de Moura, Nájla Vilar Aires</au><au>de Carvalho, Osmar Luiz Ferreira</au><au>Gomes, Roberto Arnaldo Trancoso</au><au>Guimarães, Renato Fontes</au><au>de Carvalho Júnior, Osmar Abílio</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Deep-water oil-spill monitoring and recurrence analysis in the Brazilian territory using Sentinel-1 time series and deep learning</atitle><jtitle>International journal of applied earth observation and geoinformation</jtitle><date>2022-03</date><risdate>2022</risdate><volume>107</volume><spage>102695</spage><pages>102695-</pages><artnum>102695</artnum><issn>1569-8432</issn><eissn>1872-826X</eissn><abstract>•This is the first oil-spill deep learning study in the Brazilian territory.•We compared three deep learning architectures and four backbones.•We detected hotspots using recurrence analysis from deep learning predictions.•A recurrence threshold greater than two eliminates nearly all look-alike features.•Most oil-spills events in the Campos Basin are related to oil platforms. Oil spills are a worldwide concern since they cause environmental problems and financial losses. Automatic detection plays a crucial role in rapid decision-making to reduce damage. In this context, deep learning and remote sensing are powerful tools with successful applications in many regions. However, there are still no studies on deep-water zones and in the Brazilian territory. The present research has three contributions: (1) create an oil spill dataset on the Brazilian region, (2) compare state-of-the-art deep learning models for this task, and (3) propose a novel application on the Campos Basin (Brazil's largest producer) using Sentinel-1 time series to identify more susceptible regions, encompassing 138 images from 2016 to 2021. The experts manually labeled the dataset with confirmation from the oil company and federal inspection, resulting in 800 images with 512 × 512 spatial dimensions and their respective annotations (600 for training, 125 for validation, and 75 for testing). The study compared three semantic segmentation architectures (U-net, DeepLabv3+, and LinkNet) with four backbones (ResNet-101, ResNet-50, Efficient-net-B0, and Efficient-net-B3) resulting in 12 models. The U-net with the Efficient-net-B3 backbone presented slightly better results (98% accuracy, 75% precision, 78% recall, 76% F-score, and 62% IoU). The time-series analysis considered images with 8,192 × 11,264 spatial dimensions. The entire image classification used a sliding window approach with a 128-pixel stride (averaging the overlapping pixels). Moreover, recurrence analysis detected the more susceptible areas and eliminated the look-alike. False-positive features have a high spatial presence, but with low recurrence (rarely reaching 3 events), which facilitates their elimination using a threshold recurrence. This novel analysis for oil spills is suitable for understanding the patterns along the time and providing regions for the authorities in real-time.</abstract><pub>Elsevier B.V</pub><doi>10.1016/j.jag.2022.102695</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier ISSN: 1569-8432
ispartof International journal of applied earth observation and geoinformation, 2022-03, Vol.107, p.102695, Article 102695
issn 1569-8432
1872-826X
language eng
recordid cdi_doaj_primary_oai_doaj_org_article_7df9b17c0fc14137bceebb3eb51adeb9
source Elsevier:Jisc Collections:Elsevier Read and Publish Agreement 2022-2024:Freedom Collection (Reading list)
subjects Campos Basin
Coastal management
Semantic segmentation
title Deep-water oil-spill monitoring and recurrence analysis in the Brazilian territory using Sentinel-1 time series and deep learning
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-14T12%3A14%3A17IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-elsevier_doaj_&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Deep-water%20oil-spill%20monitoring%20and%20recurrence%20analysis%20in%20the%20Brazilian%20territory%20using%20Sentinel-1%20time%20series%20and%20deep%20learning&rft.jtitle=International%20journal%20of%20applied%20earth%20observation%20and%20geoinformation&rft.au=de%20Moura,%20N%C3%A1jla%20Vilar%20Aires&rft.date=2022-03&rft.volume=107&rft.spage=102695&rft.pages=102695-&rft.artnum=102695&rft.issn=1569-8432&rft.eissn=1872-826X&rft_id=info:doi/10.1016/j.jag.2022.102695&rft_dat=%3Celsevier_doaj_%3ES0303243422000216%3C/elsevier_doaj_%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c406t-f5861ff1e03164c5825075bb4beaac62df5f3e75fda1fac08a0bb31fdfe505f33%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true