Karsten Berns

Bridging the synthetic-to-real gap in forestry perception doesn't require perfect labels: distilling knowledge from fine-grained synthetic data to coarse real-world labels significantly boosts tree instance segmentation, especially for hard-to-detect trees.

Off-road forestry environments expose critical limitations in current RGB-NIR image registration techniques like NeMAR and MURF, demanding further research into robust, multi-scale approaches.