A deep-learning framework for enhancing habitat identification based on species composition - Archive ouverte HAL
Article Dans Une Revue Applied Vegetation Science Année : 2024

A deep-learning framework for enhancing habitat identification based on species composition

1 ZENITH - Scientific Data Management
2 UMR AMAP - Botanique et Modélisation de l'Architecture des Plantes et des Végétations
3 Cirad-BIOS - Département Systèmes Biologiques
4 LIRMM | ADVANSE - ADVanced Analytics for data SciencE
5 UPVM - Université Paul-Valéry - Montpellier 3
6 SCI / MUNI - Department of Botany and Zoology [Brno]
7 Faculty of Agriculture [Belgrade]
8 Conservatoire Botanique National Méditerranéen de Porquerolles
9 Swiss Federal Institute for Forest, Snow and Landscape Research WSL
10 Department of Plant Biology and Ecology (Bilbao, Spain)
11 UNISI - Università degli Studi di Siena = University of Siena
12 UKIM - Ss. Cyril and Methodius University in Skopje
13 ZRC SAZU
14 University of Nova Gorica
15 UNIROMA - Università degli Studi di Roma "La Sapienza" = Sapienza University [Rome]
16 ZHAW - Zürcher Hochschule für Angewandte Wissenschaften = Zurich University of Applied Sciences
17 University of Bayreuth
18 ISIGE - Institut Supérieur d'Ingénierie et de Gestion de l'Environnement
19 iDiv - German Centre for Integrative Biodiversity Research
20 Universität Rostock
21 EDYSAN - Ecologie et Dynamique des Systèmes Anthropisés - UMR CNRS 7058 UPJV
22 Department of Ecoscience [Aarhus]
23 IRBio UB - Institut de Recerca de la Biodiversitat - Biodiversity Research Institute [Barcelona, Spain]
24 Institute of Botany [Kraków]
25 SAS - Slovak Academy of Sciences
26 University of Zagreb
27 UNIMOL - Università degli Studi del Molise = University of Molise
28 Uniwersytet Wroclawski = University of Wroclaw
29 Nature Research Centre [Vilnius]
30 IBER - Institute of Biodiversity and Ecosystem Research [Sofia, Bulgaria]
Milan Chytrý
Svetlana Aćić
Andraž Čarni
Florian Jansen
Remigiusz Pielech
Jozef Šibík
Zvjezdana Stančić
Alexis Joly

Résumé

Aims The accurate classification of habitats is essential for effective biodiversity conservation. The goal of this study was to harness the potential of deep learning to advance habitat identification in Europe. We aimed to develop and evaluate models capable of assigning vegetation-plot records to the habitats of the European Nature Information System (EUNIS), a widely used reference framework for European habitat types. Location The framework was designed for use in Europe and adjacent areas (e.g., Anatolia, Caucasus). Methods We leveraged deep-learning techniques, such as transformers (i.e., models with attention components able to learn contextual relations between categorical and numerical features) that we trained using spatial k-fold cross-validation (CV) on vegetation plots sourced from the European Vegetation Archive (EVA), to show that they have great potential for classifying vegetation-plot records. We tested different network architectures, feature encodings, hyperparameter tuning and noise addition strategies to identify the optimal model. We used an independent test set from the National Plant Monitoring Scheme (NPMS) to evaluate its performance and compare its results against the traditional expert systems. Results Exploration of the use of deep learning applied to species composition and plot-location criteria for habitat classification led to the development of a framework containing a wide range of models. Our selected algorithm, applied to European habitat types, significantly improved habitat classification accuracy, achieving a more than twofold improvement compared to the previous state-of-the-art (SOTA) method on an external data set, clearly outperforming expert systems. The framework is shared and maintained through a GitHub repository. Conclusions Our results demonstrate the potential benefits of the adoption of deep learning for improving the accuracy of vegetation classification. They highlight the importance of incorporating advanced technologies into habitat monitoring. These algorithms have shown to be better suited for habitat type prediction than expert systems. They push the accuracy score on a database containing hundreds of thousands of standardized presence/absence European surveys to 88.74%, as assessed by expert judgment. Finally, our results showcase that species dominance is a strong marker of ecosystems and that the exact cover abundance of the flora is not required to train neural networks with predictive performances. The framework we developed can be used by researchers and practitioners to accurately classify habitats.

Dates et versions

hal-04700157 , version 1 (17-09-2024)

Identifiants

Citer

César Leblanc, Pierre Bonnet, Maximilien Servajean, Milan Chytrý, Svetlana Aćić, et al.. A deep-learning framework for enhancing habitat identification based on species composition. Applied Vegetation Science, 2024, 27 (3), pp.e12802. ⟨10.1111/avsc.12802⟩. ⟨hal-04700157⟩
0 Consultations
0 Téléchargements

Altmetric

Partager

More