Label-Efficient 3D Forest Mapping: Self-Supervised and Transfer Learning for Instance Segmentation, Semantic Segmentation, and Species Classification
Researchers have developed a new framework for 3D forest mapping that significantly reduces the need for extensive annotated data. By employing self-supervised and transfer learning techniques, the system improves instance segmentation, semantic segmentation, and tree classification tasks. This approach not only enhances accuracy but also lowers energy consumption and carbon emissions associated with model training. AI
IMPACT Reduces data annotation burden for AI in forestry, potentially accelerating precision forestry and conservation efforts.