S2BMS - Sentinel-2 UK Butterfly Monitoring Scheme data set (Van der Plas et al., 2025, CVPR FGVC)

S2BMS (Sentinel-2 UK Butterfly Monitoring Scheme) data set

This data set consists of cloud-free Sentinel-2 (RGB+NIR, 10 m resolution) satellite images and aggregated measurements of butterfly occurrence for 1,329 locations in the UK. 

The code (to load the data) is available at https://github.com/vdplasthijs/PECL/, and in particular please see the DataSetImagePresence data loader class. The data is formatted as 1 satellite image tif file per location and 1 (geo-referenced) csv file with the species presence data for all locations.

The Sentinel-2 data was acquired from Google Earth Engine. The UKBMS data was acquired from GBIF: https://doi.org/10.15468/dl.u4rzfc

For more information, please see our upcoming paper at the 2025 CVPR FGVC12 workshop, titled:

Van der Plas, Law, Pocock, Predicting butterfly species presence from satellite imagery using soft contrastive regularisation, CVPRW proceedings (2025). 

This data is provided under a CC-BY-NC 4.0 license; please cite the above publication when using the data in publications. Thank you!