Wide-Field Imager (WFI) Time Series Dataset for Burned Area Mapping in the Brazilian Cerrado

The present dataset comprises remote sensing data collected by WFI sensors aboard the CBERS-4, CBERS-4A, and AMAZONIA-1 satellites for mapping burned areas, covering the Brazilian Cerrado in the Chapada dos Veadeiros National Park region, located in the state of Goiás, Brazil, for the years 2020, 2021, and 2022. The dataset includes information derived from the blue, green, red, and near-infrared bands, as well as from the Burned Areas Index (BAI), Enhanced Vegetation Index (EVI), Global Environmental Monitoring Index (GEMI), Normalized Difference Vegetation Index (NDVI), and Normalized Difference Water Index (NDWI).The dataset includes the following files:Spatial grid in shapefile format (.zip)Spatial grid in geopackage format (.gpkg)Dataset with spectral data (.csv)Metadata.readme (.txt)Methodology (.pdf)A script for Python using RF for supervised classification (.ipynb)Images (blue(3), green(2), red(1) and NIR(4) bands) in radiance (please, see Methodology, Section 1.5 until Lb equation) (.rar) - version 3Multispectral rasters (.tif) in top-of-atmosphere (TOA) reflectance (.zip) - version 4UsageThis dataset is intended for researchers and practitioners in machine learning and remote sensing.

It enables them to analyze the burned areas captured by the Wide Field Imager (WFI) sensor in the Brazilian Cerrado, assess the potential for mapping these areas with high temporal resolution, and develop supervised classification approaches using machine learning techniques.