Crop Classification Model

Crop Classification Model was built on the baseline of crop specific yield-data. Our commercial partner released the alpha version in 2019 which achieved an accuracy of 82-83% based on single- pixel algorithm (Sentinel-1), while the latest model released in fall 2021 reached an accuracy of 92% based on object-based Sentinel-2 algorithm, due to the increased accuracy of the field segmentation model (and harvested acres). We can detect crop types historically back to 2018 as well as in the ongoing growth season, approx. 31-37 on Zadoks growth scale depending on the region and crop-type.