Crop Rotation Modeling for Deep Learning-Based Parcel Classification from Satellite Time Series
This addresses agricultural optimization for farmers and researchers, but is incremental as it builds on existing deep learning methods for crop classification.
The paper tackles the problem of automated crop type mapping by modeling annual crop rotations with deep learning, achieving an improvement of over 6.3 mIoU points over the state-of-the-art and releasing a large-scale multi-year agricultural dataset with over 300,000 annotated parcels.
While annual crop rotations play a crucial role for agricultural optimization, they have been largely ignored for automated crop type mapping. In this paper, we take advantage of the increasing quantity of annotated satellite data to propose the first deep learning approach modeling simultaneously the inter- and intra-annual agricultural dynamics of parcel classification. Along with simple training adjustments, our model provides an improvement of over 6.3 mIoU points over the current state-of-the-art of crop classification. Furthermore, we release the first large-scale multi-year agricultural dataset with over 300,000 annotated parcels.