A constrastive semi-supervised deep learning framework for land cover classification of satellite time series with limited labels

Ienco Dino, Gaetano Raffaele, Interdonato Roberto. 2023. A constrastive semi-supervised deep learning framework for land cover classification of satellite time series with limited labels. Neurocomputing, 567:127031, 11 p.

https://doi.org/10.1016/j.neucom.2023.127031

Article de revue ; Article de recherche ; Article de revue à facteur d'impact

Version publiée - Anglais
Accès réservé aux personnels Cirad
Utilisation soumise à autorisation de l'auteur ou du Cirad.
1-s2.0-S0925231223011542-main.pdf
Télécharger (2MB) | Demander une copie

Liste HCERES des revues (en SHS) : oui

Thème(s) HCERES des revues (en SHS) : Psychologie-éthologie-ergonomie

Résumé : In this work, we present a new semi-supervised learning framework to cope with satellite image time series (SITS) classification in a data paucity scenario, considering extreme low levels of supervision. The proposed methodology, referred as SITS (Semi-Supervised Satellite Image Time Series classification method), is based on temporal convolutional neural networks and it takes advantage of both labelled and unlabelled information. SITS enforces the data to be projected in a discriminative manifold via contrastive learning, in order to produce a data representation where samples belonging to the same category are closer than the ones belonging to different ones. Pseudo-labelling is employed on unlabelled samples to take the most out of the available unlabelled information. Experiments on two study sites described by SITS of Sentinel-2 images highlight the quality of the proposed method with respect to common SITS-based classification methods and recent machine learning approaches especially tailored for the semi-supervised classification of multi-variate time series data.

Mots-clés Agrovoc : satellite d'observation de la Terre, apprentissage machine, classification, cartographie de l'occupation du sol, méthodologie, classification des terres, imagerie par satellite, réseau de neurones, couverture du sol

Mots-clés libres : Machine Learning, Deep Learning, Satellite Image Time Series, Land cover classification, Time series

Classification Agris : U30 - Méthodes de recherche
U10 - Informatique, mathématiques et statistiques

Champ stratégique Cirad : CTS 5 (2019-) - Territoires

Auteurs et affiliations