Canopy Height Estimation in Brazilian Amazon using Multisource Remote Sensing Datasets

Forests are important carbon sinks that help mitigate climate change and maintain ecological balance and biodiversity. Accurate measurement of forest parameters, such as canopy height and aboveground biomass, is essential for understanding their impact on the global carbon cycle and for effective land-use planning. This study focuses on estimating canopy height in the Brazilian Amazon to improve understanding of forest ecosystem health by integrating multisource remote sensing datasets from Sentinel-1, Sentinel-2, ALOS-2/PALSAR-2, and GEDI and evaluating the performance of three deep learning architectures: a ResNet-based, α-2MDU and Hy-Tec models. Our results show that α-2MDU model has a better performance.