Researchers have developed a new framework to expand tree inventories using scarce annotations from UAV imagery. This method iteratively self-trains a DeepLabV3 network, combining semantic segmentation with crown-distance regression. The framework filters candidate crowns based on confidence, class probability, distance map, and morphological characteristics before adding them to the training set, ensuring consistency with the reference inventory. Evaluations in the Amazon rainforest and Cerrado showed significant increases in retained crowns and improved crown-level F1 scores compared to baseline models, offering a cost-effective approach for species inventory expansion. AI
IMPACT This research offers a more efficient method for ecological data collection, potentially aiding conservation efforts and climate modeling.
RANK_REASON The cluster contains an academic paper detailing a new methodology and its evaluation. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →