One of the responsibilities of the Common Agricultural Policy (CAP) is to control the agricultural activity compliance with the specific requirements in order to increase operational efficiency and to prevent possible irregularities. In the EU, the Member States’ Land Parcel Identification Systems (LPIS), which provide detailed digital geometries of agricultural reference parcels to aid the management of the CAP, are being released as open access data in an increasing number of regions. Farmers use the LPIS to declare their cropping practices, including specific environmental measures where relevant, in an annual aid application. The combined use of detailed reference parcel databases with satellite data facilitates near-real-time information gathering for a large number of agricultural parcels.

However, in-situ data also need to be collected in order to train and validate the information extraction process. This can be challenging as data may not be available until late in the growing season and may also not be considered as the ground-truth. Furthermore, traditional in-situ ground-truth collection lacks the scale and possibility for automated integration into big data analyses and is prone to sampling errors.

CALLISTO will develop an in-season availability of parcel-level crop type information using Sentinel-1 and Sentinel-2 data, aiming to identify across a large area, parcels that potentially do not comply with CAP rules that are then targeted for inspection during a one-day field survey with Unmanned Aerial Vehicle (UAV) (smart sampling). CALLISTO will make use of Deep Learning to perform large-scale parcel-level crop classification and outlier detection for grassland. The trained models will then be loaded on the UAV mounted machine to produce near-real time inference at the edge. We will then evaluate the feasibility of this approach and its ability to efficiently scale across large areas along with its implications for CAP.

This pilot use case works towards the exhaustive monitoring of the CAP (CAP post 2020), which will be accomplished using the collection and analysis of several types of heterogeneous and large data collections, such as UAV imagery, geo-tagged photos and street-level imagery from vehicles. Artificial Intelligence (AI) technologies are also involved in the data analysis in order to support EU implementation bodies and policymakers in EU CAP monitoring.