5.46 million image patches of potential deforestation labeled by the Amazonia People Environment Institute-Imazon. The data ranges from January 2021 to July 2025 and is hosted by Harvard Dataverse.
Use Cases
- Training deforestation detection models based on satellite imagery patches.
- Benchmarking hybrid human-AI model fusion techniques for land use change.
- Analyzing spatial and temporal patterns of deforestation in the Brazilian Amazon.
Strengths
- 5.46 million labeled image patches provide a substantial sample size.
- Data covers a multi-year period from January 2021 to July 2025.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
Provenance
- Source
- Amazonia People Environment Institute-Imazon
- Collection Method
- Patches of potential deforestation were labeled, likely from satellite imagery.
- Time Range
- January 2021 to July 2025
- Freshness
- Last updated 2026-05-29 11:35:13
- Geography
- Brazilian Legal Amazon