Forest Structure

Forest structure research focuses on accurately mapping and quantifying various aspects of forest composition and geometry, primarily to improve forest monitoring and management. Current efforts leverage advanced machine learning techniques, particularly deep learning models like neural radiance fields and UNets, often incorporating transfer learning strategies and deep ensembles, to process multi-source satellite imagery (optical and SAR) for high-resolution, large-scale forest structure estimations. This work is crucial for addressing challenges like deforestation and climate change by providing accurate, scalable data on key forest variables such as height, density, and canopy cover, improving the reliability of forest inventories and informing conservation efforts.