Assessment of pixel-oriented k-NN machine learning algorithm performance for the interannual remote sensing monitoring of eelgrass beds at the mouth of the Romaine

Eelgrass cover extent is among the most reliable indicators for measuring changes in coastal ecosystems. Eelgrass has colonized the mouth of the Romaine River and has become a part of environmental monitoring there since 2013. The presence of eelgrass in this area is an essential factor for the earl...

Full description

Saved in:

Bibliographic Details
Published in:	Environmental monitoring and assessment Vol. 195; no. 8; p. 939
Main Authors:	Lemieux, P., Lalumière, C., Fugaru, N., Gilbert, J.-P., Tremblay, A.
Format:	Journal Article
Language:	English
Published:	Cham Springer International Publishing 01-08-2023 Springer Nature B.V
Subjects:	Algorithms Aquatic plants Atmospheric Protection/Air Quality Control/Air Pollution Change detection Coastal ecosystems Data acquisition Data collection Detection Earth and Environmental Science Ecology Ecosystem Ecosystems Ecotoxicology Edge detection Environment Environmental Management Environmental Monitoring Environmental science Machine Learning Monitoring Monitoring/Environmental Analysis Pixels Remote monitoring Remote sensing Remote Sensing Technology Review Sea grasses Training Workflow Zosteraceae K-NN Eelgrass Pixel oriented Classification Machine learning
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Eelgrass cover extent is among the most reliable indicators for measuring changes in coastal ecosystems. Eelgrass has colonized the mouth of the Romaine River and has become a part of environmental monitoring there since 2013. The presence of eelgrass in this area is an essential factor for the early detection of changes in the Romaine coastal ecosystem. This will act as a trigger for an appropriate environmental response to preserve ecosystem health. In this paper, a cost- and time-efficient workflow for such spatial monitoring is proposed using a pixel-oriented k-NN algorithm. It can then be applied to multiple modellers to efficiently map the eelgrass cover. Training data were collected to define key variables for segmentation and k-NN classification, providing greater edge detection for the presence of eelgrass. The study highlights that remote sensing and training data must be acquired under similar conditions, replicating methodologies for collecting data on the ground. Similar approaches must be used for the zonal statistic requirements of the monitoring area. This will allow a more accurate and reliable assessment of eelgrass beds over time. An overall accuracy of over 90% was achieved for eelgrass detection for each year of monitoring.
Bibliography:	ObjectType-Article-2 SourceType-Scholarly Journals-1 ObjectType-Feature-3 content type line 23 ObjectType-Review-1
ISSN:	0167-6369 1573-2959
DOI:	10.1007/s10661-023-11468-3