Provided is an automatic drawing method of high-resolution global burned areas based on time-sensitive spectral index, which can effectively distinguish the bare land, terrain shadows, unrecovered old burned area and other non-burned areas that are easily confused with the burned areas in the case that the medium and high resolution satellite time series images are discontinuous, and provides a powerful tool for the automatic extraction of high-resolution burned areas; combining the designed time-sensitive spectral index, using the global sample library of burned and non-burned areas and the machine learning method based on ecosystem partition, the automatic mapping of high-resolution global burned areas is realized.
展开▼