Assessing the repeatability of automated seafloor classification algorithms, with application in marine protected area monitoring
Assessing the repeatability of automated seafloor classification algorithms, with application in marine protected area monitoring
The number and areal extent of marine protected areas worldwide is rapidly increasing as a result of numerous national targets that aim to see up to 30% of their waters protected by 2030. Automated seabed classification algorithms are arising as faster and objective methods to generate benthic habitat maps to monitor these areas. However, no study has yet systematically compared their repeatability. Here we aim to address that problem by comparing the repeatability of maps derived from acoustic datasets collected on consecutive days using three automated seafloor classification algorithms: (1) Random Forest (RF), (2) K–Nearest Neighbour (KNN) and (3) K means (KMEANS). The most robust and repeatable approach is then used to evaluate the change in seafloor habitats between 2012 and 2015 within the Greater Haig Fras Marine Conservation Zone, Celtic Sea, UK. Our results demonstrate that only RF and KNN provide statistically repeatable maps, with 60.3% and 47.2% agreement between consecutive days. Additionally, this study suggests that in low-relief areas, bathymetric derivatives are non-essential input parameters, while backscatter textural features, in particular Grey Level Co-occurrence Matrices, are substantially more effective in the detection of different habitats. Habitat persistence in the test area between 2012 and 2015 was 48.8%, with swapping of habitats driving the changes in 38.2% of the area. Overall, this study highlights the importance of investigating the repeatability of automated seafloor classification methods before they can be fully used in the monitoring of benthic habitats.
Automated seafloor classification, Autonomous underwater vehicles, Benthic habitat maps, Grey level co-occurrence matrices, Machine learning algorithms, Sidescan sonar
Zelada Leon, America
f0d8abf3-3b85-44dd-b466-f36eb1791e48
Huvenne, Veerle A.i.
f22be3e2-708c-491b-b985-a438470fa053
Benoist, Noëlie M.a.
28b7771d-bfa6-4a2a-8426-bb3c24a5cccd
Ferguson, Matthew
5eaece6b-ba5c-4351-b402-f21016f24aaa
Bett, Brian J.
61342990-13be-45ae-9f5c-9540114335d9
Wynn, Russell B.
72ccd765-9240-45f8-9951-4552b497475a
15 May 2020
Zelada Leon, America
f0d8abf3-3b85-44dd-b466-f36eb1791e48
Huvenne, Veerle A.i.
f22be3e2-708c-491b-b985-a438470fa053
Benoist, Noëlie M.a.
28b7771d-bfa6-4a2a-8426-bb3c24a5cccd
Ferguson, Matthew
5eaece6b-ba5c-4351-b402-f21016f24aaa
Bett, Brian J.
61342990-13be-45ae-9f5c-9540114335d9
Wynn, Russell B.
72ccd765-9240-45f8-9951-4552b497475a
Zelada Leon, America, Huvenne, Veerle A.i., Benoist, Noëlie M.a., Ferguson, Matthew, Bett, Brian J. and Wynn, Russell B.
(2020)
Assessing the repeatability of automated seafloor classification algorithms, with application in marine protected area monitoring.
Remote Sensing, 12 (10), [1572].
(doi:10.3390/rs12101572).
Abstract
The number and areal extent of marine protected areas worldwide is rapidly increasing as a result of numerous national targets that aim to see up to 30% of their waters protected by 2030. Automated seabed classification algorithms are arising as faster and objective methods to generate benthic habitat maps to monitor these areas. However, no study has yet systematically compared their repeatability. Here we aim to address that problem by comparing the repeatability of maps derived from acoustic datasets collected on consecutive days using three automated seafloor classification algorithms: (1) Random Forest (RF), (2) K–Nearest Neighbour (KNN) and (3) K means (KMEANS). The most robust and repeatable approach is then used to evaluate the change in seafloor habitats between 2012 and 2015 within the Greater Haig Fras Marine Conservation Zone, Celtic Sea, UK. Our results demonstrate that only RF and KNN provide statistically repeatable maps, with 60.3% and 47.2% agreement between consecutive days. Additionally, this study suggests that in low-relief areas, bathymetric derivatives are non-essential input parameters, while backscatter textural features, in particular Grey Level Co-occurrence Matrices, are substantially more effective in the detection of different habitats. Habitat persistence in the test area between 2012 and 2015 was 48.8%, with swapping of habitats driving the changes in 38.2% of the area. Overall, this study highlights the importance of investigating the repeatability of automated seafloor classification methods before they can be fully used in the monitoring of benthic habitats.
Text
remotesensing-12-01572-v2
- Version of Record
More information
Accepted/In Press date: 11 May 2020
Published date: 15 May 2020
Additional Information:
Funding Information:
Funding: This research was funded by the NERC Marine Environmental Mapping Programme (MAREMAP), the DEFRA project “Investigating the feasibility of using AUV and Glider technology for mapping and monitoring of the UK MPA network (MB0118)”, and the NERC Climate Linked Atlantic Sector Science (CLASS) project (Grant no: NE/R015953/1); A.Z.L. was funded by the CONICYT PFCHA/MAGISTER BECAS CHILE/2017—73180206 Fellowship program.
Publisher Copyright:
© 2020 by the authors.
Keywords:
Automated seafloor classification, Autonomous underwater vehicles, Benthic habitat maps, Grey level co-occurrence matrices, Machine learning algorithms, Sidescan sonar
Identifiers
Local EPrints ID: 444442
URI: http://eprints.soton.ac.uk/id/eprint/444442
ISSN: 2072-4292
PURE UUID: 57123342-daf1-4079-ad4d-04a740c19fbc
Catalogue record
Date deposited: 19 Oct 2020 16:33
Last modified: 06 Jun 2024 01:42
Export record
Altmetrics
Contributors
Author:
America Zelada Leon
Author:
Veerle A.i. Huvenne
Author:
Noëlie M.a. Benoist
Author:
Matthew Ferguson
Author:
Brian J. Bett
Author:
Russell B. Wynn
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics