KlimBa
Deep learning for the recognition of individual trees in aerial images
Deep learning methods are used to recognise individual trees in digital orthophotos with 20 cm ground resolution. The background is the generation of reference data for the classification of satellite data from the Copernicus mission in the KlimBa project.
The deep learning networks are trained with field recordings of the federal forest inventory and manually digitised tree crowns. Research focuses on fine-tuning the networks to extract the tree crowns, matching the extracted tree crowns with the tree coordinates measured in the field, assigning tree species and developing a validation concept to assess the quality of the network.
More information
Further information on this project can also be found on the following external pages:
