Large Scale Mowing Event Detection on Dense Time Series Data Using Deep Learning Methods and Knowledge Distillation

Cultivated grasslands are of great importance to national strategies and in synergy with Common Agricultural Policy(CAP) there is an emerging demand for new innovative methods capable of monitoring vegetation in large areas over growing periods. Vegetation of this kind produces a large volume of biomass that is economically valuable for various uses. Deep learning methods have shown their potential in combination with Earth Observation data, but data availability for mowing detection task remains low. To combat the lack of available data, we developed this dataset for three different Regions of Interest in Greece that contains over 1600 different parcels. This dataset is part of the paper 'Large Scale Mowing Event Detection on Dense Time Series Data Using Deep Learning Methods and Knowledge Distillation' published in ISPRS Archives (EARSEL 44th Symposium 2025).