Integrating Deep Learning and UAV Multispectral Data for Improved Estimation of Grassland Aboveground Biomass in Arid Regions

Accurate estimation of grassland aboveground biomass (AGB) is essential for understanding ecosystem productivity, evaluating grassland degradation, and supporting sustainable grassland management. However, in arid and semi-arid regions, the heterogeneous background of vegetation and bare soil, together with the spatial scale mismatch between ground observations and satellite pixels, poses significant challenges for accurate AGB estimation. This study proposes a ground–UAV–satellite cross-scale collaborative framework that integrates UAV multispectral imagery, vegetation classification, and deep learning algorithms to improve grassland AGB estimation accuracy.