Automated Satellite-Based Monitoring of Agricultural Subsidies in Estonia

Organisation type

Public administration at national level

Use case area

Agriculture, fisheries, forestry and food
Science and technology

Geographic scope

Estonia

Domain

Economy
Environment

Challenge:

The Estonian Agricultural Registers and Information Board (ARIB) faces the complex task of verifying subsidy claims by farmers, ensuring compliance with the European Common Agricultural Policy (CAP). One of the conditions under the CAP that is frequently violated on less intensively used agricultural lands is a lack of obligatory mowing. Historically, the checks relied on manual field inspections, which were costly, time-consuming, and covered only a small fraction of the land. With the rising costs of human resources and the vast area needing inspection, ARIB required a more efficient and accurate method to ensure that subsidies were allocated fairly and correctly.

Approach:

Faced with costly and inefficient manual field inspections, ARIB partnered with Tartu Observatory and CGI Estonia to create the SATIKAS system. Named after the Estonian phrase for "a system that uses satellite data", SATIKAS utilises Sentinel-1 (S1) radar and Sentinel-2 (S2) optical satellite imagery to monitor farmers' compliance with the mowing requirement. The system operates annually from May to October, covering the entire growing season and detecting all mowing activities on agricultural grasslands in Estonia. As the optical measurements of S2 are prone to failure in cloudier months, the radar equipment of S1, which penetrates clouds easily, offers a chance to fill in those gaps. Each season, over 100 radar images and more than 50 optical images are used, along with the reference data of farmer fields, historical inspection logs and meteorological data from the Estonian Weather Service. The original satellite pictures are converted into several processed data layers. After that, the average value of each date’s image and each grassland is calculated to form the time series. Then, a deep learning algorithm is used on that time series to detect the mowing information. Mowing and its range can be observed from sudden changes in the biomass levels in the time series parameters of the satellite data.

Outcomes:

The SATIKAS system has significantly reduced the need for expensive and labour-intensive field visits by automating the detection of mowing activities through satellite data. The initiative has increased inspection coverage, resulting in a €500,000 reduction in incorrect subsidy payments annually in Estonia. Moreover, the system has enhanced the efficiency of inspectors by directing them only to non-compliant fields, thereby improving resource allocation. This data-driven initiative not only ensures compliance with the European Common Agricultural Policy (CAP) but also fosters a transparent and efficient subsidy distribution process. The combination of satellite data, deep learning techniques, and real-time reporting illustrates how data-driven insights can be achieved and sets new standard for accuracy and efficiency in the field.

Data Sources:

Several data sources are employed, these include Sentinel-1 radar images and Sentinel-2 optical satellite images, which provide detailed observations from space. Additionally, historical field inspection logs are used to incorporate past observations and records. Meteorological data from the Estonian Weather Service is also integrated, offering valuable insights into weather patterns and conditions relevant to the analysis.

Additional Information:

Point of Contact:

priapria [dot] ee (pria[at]pria[dot]ee)