Dataset : Tree species classification at the pixel-level using deep learning and multispectral time series in an imbalanced context

This dataset contains 7 tree species classification maps generated as part of the research presented in the journal paper:  Florian Mouret, David Morin, Milena Planells, Cécile Vincent-Barbaroux. Tree species classification at the pixel-level using deep learning and multispectral time series in an imbalanced context. https://doi.org/10.48550/arXiv.2408.08887

The maps, stored in GeoTIFF format, represent pixel-level classifications of 10 tree species in Centre-Val de Loire, France. They were derived from Sentinel-2 time series acquired between 2019 and 2020. The maps have a spatial resolution of 10 meters. These maps illustrate the performance of the Random Forest algorithm and different deep learning models in classifying tree species based on multispectral time series data. The analysis focuses on the problem of imbalanced data, where the training data is largely dominated by oak. They are derived from the model outputs discussed in the associated journal paper. 

The maps were generated using the Python library iota2: https://framagit.org/fl.mouret/tree_species_classification_iota2

Keywords: tree species classification, deep learning, multispectral, remote sensing, time series, pixel-level, imbalanced data.