Wetland Mapping and Land Surface Temperature Inversion Product in Beijing Produced with SDGSAT-1 Data

This data was developed by the Wetland remote sensing research team of Aerospace Information Research Institute, Chinese Academy of Sciences. Based on the spectral features and index features obtained from six SDGSAT-1 multispectral images acquired in February, March, and April 2022, we carried out wetland mapping in Beijing using the GBDT (Gradient Boosting Decision Tree). The spatial resolution is 10m, overall accuracy is 86.99%, Kappa coefficient is 0.85, Macro-F1 is 0.86.

The meaning of the attribute value is: 1 Water, 2 Marsh, 3 Paddy field, 4 Impervious surface, 5 Forest and Scrub, 6 Grassland, 7 Farmland. The LST product is based on two SDGSAT-1 TIR images on April 1st, the spatial resolution is 30m, in degrees Celsius.