Supervised and self-supervised land-cover segmentation & classification of the Biesbosch wetlands

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Meijling, Eva Gmelich, Del Prete, Roberto, Visser, Arnoud
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913861484937216
author Meijling, Eva Gmelich
Del Prete, Roberto
Visser, Arnoud
author_facet Meijling, Eva Gmelich
Del Prete, Roberto
Visser, Arnoud
contents Accurate wetland land-cover classification is essential for environmental monitoring, biodiversity assessment, and sustainable ecosystem management. However, the scarcity of annotated data, especially for high-resolution satellite imagery, poses a significant challenge for supervised learning approaches. To tackle this issue, this study presents a methodology for wetland land-cover segmentation and classification that adopts both supervised and self-supervised learning (SSL). We train a U-Net model from scratch on Sentinel-2 imagery across six wetland regions in the Netherlands, achieving a baseline model accuracy of 85.26%. Addressing the limited availability of labeled data, the results show that SSL pretraining with an autoencoder can improve accuracy, especially for the high-resolution imagery where it is more difficult to obtain labeled data, reaching an accuracy of 88.23%. Furthermore, we introduce a framework to scale manually annotated high-resolution labels to medium-resolution inputs. While the quantitative performance between resolutions is comparable, high-resolution imagery provides significantly sharper segmentation boundaries and finer spatial detail. As part of this work, we also contribute a curated Sentinel-2 dataset with Dynamic World labels, tailored for wetland classification tasks and made publicly available.
format Preprint
id arxiv_https___arxiv_org_abs_2505_21269
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Supervised and self-supervised land-cover segmentation & classification of the Biesbosch wetlands
Meijling, Eva Gmelich
Del Prete, Roberto
Visser, Arnoud
Computer Vision and Pattern Recognition
Image and Video Processing
68
I.4.6
Accurate wetland land-cover classification is essential for environmental monitoring, biodiversity assessment, and sustainable ecosystem management. However, the scarcity of annotated data, especially for high-resolution satellite imagery, poses a significant challenge for supervised learning approaches. To tackle this issue, this study presents a methodology for wetland land-cover segmentation and classification that adopts both supervised and self-supervised learning (SSL). We train a U-Net model from scratch on Sentinel-2 imagery across six wetland regions in the Netherlands, achieving a baseline model accuracy of 85.26%. Addressing the limited availability of labeled data, the results show that SSL pretraining with an autoencoder can improve accuracy, especially for the high-resolution imagery where it is more difficult to obtain labeled data, reaching an accuracy of 88.23%. Furthermore, we introduce a framework to scale manually annotated high-resolution labels to medium-resolution inputs. While the quantitative performance between resolutions is comparable, high-resolution imagery provides significantly sharper segmentation boundaries and finer spatial detail. As part of this work, we also contribute a curated Sentinel-2 dataset with Dynamic World labels, tailored for wetland classification tasks and made publicly available.
title Supervised and self-supervised land-cover segmentation & classification of the Biesbosch wetlands
topic Computer Vision and Pattern Recognition
Image and Video Processing
68
I.4.6
url https://arxiv.org/abs/2505.21269