A joint research team from Beijing Forestry University, Manchester Metropolitan University, and Tsinghua University has developed an artificial intelligence model that can estimate tree heights from standard RGB satellite images with sub-meter accuracy, offering a low-cost alternative to lidar-based forest monitoring. Published in the Journal of Remote Sensing on October 20, 2025, the study introduces a framework that combines large vision foundation models (LVFMs) with self-supervised learning to produce high-resolution canopy height maps.
Forests and plantations are critical for carbon sequestration, but accurately monitoring their growth has traditionally required expensive lidar surveys or labor-intensive field measurements. The new model, tested in the Fangshan District of Beijing and the Saihanba forest, achieved a mean absolute error of 0.09 meters and an R² of 0.78 when compared with airborne lidar data, outperforming conventional deep learning methods. It also demonstrated over 90% accuracy in single-tree detection and strong correlations with above-ground biomass measurements.
The model uses the DINOv2 large vision foundation model as a feature extractor, combined with a self-supervised feature enhancement unit and a lightweight convolutional height estimator. By using one-meter-resolution Google Earth imagery as input and UAV-based lidar data for training, the AI produced canopy height maps that closely matched ground truth. The approach significantly outperformed global canopy height products, capturing subtle variations in tree crown structure often missed by existing models.
“Our model demonstrates that large vision foundation models can fundamentally transform forestry monitoring,” said Dr. Xin Zhang, corresponding author at Manchester Metropolitan University. “By combining global image pretraining with local self-supervised enhancement, we achieved lidar-level precision using ordinary RGB imagery. This approach drastically reduces costs and expands access to accurate forest data for carbon accounting and environmental management.”
The ability to reconstruct annual growth trends from archived satellite imagery provides a scalable solution for long-term carbon sink monitoring and precision forestry management. The model’s cross-regional adaptability was confirmed when applied to the geographically distinct Saihanba forest, maintaining robust accuracy. This innovation bridges the gap between expensive lidar surveys and low-resolution optical methods, enabling detailed forest assessment with minimal data requirements.
The AI-based mapping framework offers a powerful and affordable approach for tracking forest growth, optimizing plantation management, and verifying carbon credits under initiatives such as China's Certified Emission Reduction program. Its adaptability across ecosystems makes it suitable for global afforestation and reforestation monitoring programs. Future research will extend this method to natural and mixed forests, integrate automated species classification, and support real-time carbon monitoring platforms. As the world advances toward net-zero goals, such intelligent, scalable mapping tools could play a central role in achieving sustainable forestry and climate-change mitigation.
For more details, see the original study at https://spj.science.org/doi/10.34133/remotesensing.0880.
