Multi-features Deep Learning Framework for aboveground biomass mapping in semi-arid forests by integrating LiDAR with Sentinel-1 and Sentinel-2 time series

We have provided the data and code for this research here.We have provided all the variables of Sentinel 1, including polarization metrics and texture features, which can be queried in a file named "S1". All variables of Sentinel 2, including band, vegetation index, biophysical features and seasonal features, can be queried in files named "S2_band", "S2_VI", "S2_Biophysical" and "S2_Seasonal_features". The GEDI data of the study area from 2019 to 2023 that we obtained can also be queried in this database.The measured data is also provided and it is stored in the "measured_data" file.In addition, we have provided all the codes for implementing this research, including the codes for all the models used in this study, as well as the codes for processing Sentinel 1 data, Sentinel 2 data, and GEDI data to obtain the feature data required for this study.These codes are stored in the file "code".