This data set is about the forest cover data of southeast China in 2018. Combining the characteristics and advantages of Landsat and sentinel-2 remote sensing images, spectral, spatial and temporal feature sets reflecting different forest types were studied and established respectively. The characteristics of Landsat long time series and time series harmonic analysis technique were utilized to establish the time feature set of forest type extraction. Based on the spectral-spatial-temporal feature set, the main features in different regions were studied and established by using the Random Forest-Recursive feature elimination algorithm with the support of reference data. According to the spectral-spatial-temporal feature set of different regions, using four machine learning algorithms to establish forest type classification model. Then the forest type maps with 10 spatial resolution in 2018 for southeast China were generated using the determined best fit model for different regions.
Comment list ( 0 )