Name in original language
Algoritmo Adetop de Inteligencia Artificial para la Supervisión Forestal
Initiative overview
Peru's Osinfor has deployed an updated version of the Adetop algorithm, an AI system that uses multispectral satellite imagery from the European Space Agency's Sentinel-2 satellite and NASA's Landsat constellation to detect concentrated tree extraction across Peruvian territory. The algorithm was trained using drone imagery and field supervision materials collected by Osinfor, as well as historical satellite imagery dating back to 2019. It cross-references detections with data on authorised concessions and protected areas. According to researchers from the Universities of Sheffield and Cambridge, who travelled to Madre de Dios to support model validation, the system achieves over 90% reliability for detecting selective logging, with precision five times higher than the previous version. The system also selects cloud-free satellite images to reduce interpretation errors.
All data is exportable in GEOJSON vector format, streamlining administrative reporting and enforcement proceedings. Osinfor collaborates with the Ministerio Público to ensure that data produced by the system is admissible as evidence in criminal investigations. The initiative also involves technology transfer to the regional governments of Madre de Dios, Ucayali, and Loreto, with regional Monitoring Centres expected to be established shortly to expand supervisory coverage and enable faster response.
