A burned area mapping method for the FY-3D MERSI based on the single-temporal L1 data and multi-temporal daily active fire products

Wildfire occurs throughout the world, resulting in global ecology, society, and economic impacts. Burned areas are important for fire monitoring. In order to develop the capability of the Fengyun (FY) satellite to detect burned areas, a burned area mapping method based on the Fengyun-3D satellite (FY-3D) Medium Resolution Spectral Imager (MERSI) L1 data and daily active fire products is presented in this paper. Referring to other methods based on the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Advanced Very High-Resolution Radiometer (AVHRR), active fires based on daily active fire products are selected and clustered as burned samples. Two indices are used as spectral features of burned areas based on the single temporal FY-3D MERSI 250-m post-fire data. By taking advantage of spatial and spectral features for land cover in each burned cluster, burned samples are extended to identify burned areas by a region growing algorithm. The method is applied in two study areas and the result is compared with the Collection 6 MODIS burned area products (MCD64A1) based on the reference values obtained by the human-machine interface for validation. The results show that this method has a relatively high correspondence with the reference values, and its accuracy is higher than that of the MCD64A1. Certainly, this method inevitably has some limitations. But overall, the method can be used to extract burned areas by using the FY-3D MERSI data quickly and precisely, especially over large areas.