Early Classification for Agricultural Monitoring from Satellite Time Series
This work addresses agricultural monitoring for farmers and authorities by enabling earlier crop type identification, though it is incremental as it builds on existing classification models.
The paper tackled the problem of early crop type classification from satellite time series by introducing an end-to-end trainable early classification mechanism that augments existing models with a stopping probability based on observed data. The results showed that a recurrent neural network with this mechanism could distinguish many crop types before the end of the vegetative period, with classification times linked to plant phenology events.
In this work, we introduce a recently developed early classification mechanism to satellite-based agricultural monitoring. It augments existing classification models by an additional stopping probability based on the previously seen information. This mechanism is end-to-end trainable and derives its stopping decision solely from the observed satellite data. We show results on field parcels in central Europe where sufficient ground truth data is available for an empiric evaluation of the results with local phenological information obtained from authorities. We observe that the recurrent neural network outfitted with this early classification mechanism was able to distinguish the many of the crop types before the end of the vegetative period. Further, we associated these stopping times with evaluated ground truth information and saw that the times of classification were related to characteristic events of the observed plants' phenology.