Correctly delineated agricultural parcels are often mandatory for agriculture-related downstream applications. For instance in applications related to CAP, the delineated boundaries can aid the farmers speed up the declaration process and the paying agencies to better monitor changes in agricultural use. In the example shown here, we have used Sentinel-2 data from Sentinel-Hub services for the month of June 2020 over Lithuania to produce automatically detect (agricultural) parcels.
The parcels are predicted using a modified version of the ResUnet-a deep learning architecture, trained on GSAA data. The parcels are colour-coded according to their area in hectares, which is displayed when clicking on a parcel. Adjust the opacity of the fill colour and zoom in to assess the quality of the delineation with respect to most cloudless observation from Sentinel-2, shown in the background.
More details about the workflow can be found in this blog-post. For more info about automatic parcel boundary detection and delineation contact us at eoresearch _at_ sinergise dot com.