Repository navigation
DeepBiomass is a novel deep-learning method that allows Deep Learning for biomass estimation from point cloud data.
In the first step (Stage 1), it expects a number of (las-format) point clouds and a biomass raster as input.
In the second step (Stage 2), an MLP can be trained to predict a feature vector derived for one epoch of biomass to a second epoch (different acquisition, other sensor, ...).