Research project Monitoring of Environmental Practices for Sustainable Agriculture Supported by Earth Observation
General introduction
Continuous monitoring of agricultural management activities on a large scale, with the help of satellite imagery, is a step closer through the work of the Horizon2020 project ENVISION. The researchers developed and tested innovative Earth Observation (EO)-based and Machine Learning (ML) technologies. The capacity to monitor agricultural management automatically and reliably is very important to achieve the agri-environmental objectives of the Common Agricultural Policy (CAP). ILVO's task in ENVISION was situated around cloud services enabling large-scale monitoring of soil quality through satellite data and AI technology. Flanders was the test area: a region-wide map of organic carbon in the topsoil was developed for this region. Such a map proves to offer valuable insights to policy authorities for strategic planning. It also becomes possible to intervene in (future) soil health with targeted actions.
Research approach
ILVO's used a 5-step methodology to estimate soil quality at regional level. We utilized an extensive image collection covering more than 8,000 images across 680,000 hectares of agricultural land. Soil quality maps at both pixel and parcel levels were created, providing detailed classifications of SOC (Soil Organic Carbon) values. These maps enable targeted soil health interventions and offer valuable insights to policy authorities for strategic planning. They classify SOC values in relation to average conditions while considering soil-pedoclimatic factors.
Relevance/Valorization
ILVO aims to valorize the ENVISION results by utilizing the developed cloud services for different target groups at both regional and European levels (Envsion platform). At the regional level, a developed application is available, allowing Flemish farmers to view their agricultural parcels on a map. For each parcel, it is possible to see the internal variability of the topsoil organic carbon as estimated using satellite data and the AI model. The software application and its Application Programming Interface (API) were developed by ILVO, and the results were demonstrated at the AgriTEF event 2023, in collaboration with the Flemish Department of Agriculture and Fisheries.
