Developed by the Remote Sensing and GIS Lab (LAPIG)
This repository contains scripts and resources for the mapping of pasturelands in the Brazil, as part of the MapBiomas Brazil initiative. The mapping is based on multi-temporal remote sensing imagery from the Landsat (30m) and Sentinel (10m) satellite programs.
The classification processes include data preparation, model training, classification, and a series of post-processing routines to ensure temporal and spatial consistency of the maps.
For detailed methodology and technical specifications, refer to the Pasture Appendix of the Algorithm Theoretical Basis Document (ATBD).
The repository is organized into subfolders by image source and processing resolution, following the MapBiomas classification workflow for the pasturelands:
-
lulc_30m_landsat:
Scripts for generating annual LULC maps at 30-meter resolution, using Landsat imagery. -
lulc_10m_sentinel:
Scripts for generating annual LULC maps at 10-meter resolution, using Sentinel-2 imagery. -
product_pasture_vigor_condition_30m_MOD13Q1_EVI:
Scripts for generating Pasture Vigor Condition product at 30-meter resolution, using MOD13Q1 EVI and LULC 30m pasture maps. -
product_pasture_biomass_productivity_30m_GPW_uGPP:
Scripts for generating Pasture Biomass Productivity product at 30-meter resolution, using GPW uncalibrated Gross Primary Productivity (uGPP) and LULC 30m pasture maps.
Each subfolder includes a step-by-step processing chain with processing/classification scripts, filtering procedures, and additional assets used in the generation of MapBiomas Pasture LULC collections and products.
If you use any part of this repository or the resulting data in your work, please cite MapBiomas and LAPIG accordingly. For official data access and citation guidelines, visit the MapBiomas Terms of Use. The MapBiomas data are public, open and free under Creative Commons CC-BY license.
For questions, suggestions, or to report issues, please contact: