Accurate tree counts are essential for restoration, yet conventional methods struggle across heterogeneous landscapes. Researchers developed a deep-learning framework combining multisensor satellite imagery to estimate tree abundance. Applied to Rwanda and Tanzania, it counted 5.25 billion trees and showed that more than half occur outside areas classified as tree cover.
Global tree-planting pledges have increased sharply since the Paris Agreement, creating demand for reliable baselines showing how many trees already exist and where they occur. Field inventories are costly and spatially limited, while repeated acquisition of submeter satellite or Unmanned Aerial Vehicle (UAV) imagery is difficult to scale. Object-detection methods also struggle when crowns overlap, occupy fractions of pixels, or lack clear boundaries. Coarser approaches based on vegetation indices, climate, or topography may smooth local variation and underrepresent trees in nonforest systems. Given these challenges, in-depth research is needed to develop scalable, affordable, robust, observation-based methods for counting trees across heterogeneous natural landscapes.
A research team led by the University of Copenhagen published (DOI: 10.34133/remotesensing.0970) the study on August 18, 2026, in the J ournal of R emote S ensing . The team developed a multisensor deep-learning framework using PlanetScope, Sentinel-1, and Sentinel-2 imagery. Instead of detecting every crown individually, the system estimates tree numbers from spatial, spectral, and textural patterns within image patches. This approach targets a restoration-monitoring challenge: producing consistent, spatially detailed tree-count baselines across forests and nonforest landscapes without relying on expensive, repeatedly acquired submeter imagery.
The framework estimated about 5.25 billion trees across Rwanda and Tanzania: roughly 213 million in Rwanda and 5.04 billion in Tanzania after calibration. Only 48.8% fell within the WorldCover "tree cover" class; the other 51.2% occurred across shrublands, grasslands, croplands, built-up areas, wetlands, sparse vegetation, and mangroves. Unlike forest-focused inventories or coarse statistical extrapolations, the approach captures trees across heterogeneous landscapes. Its image-level regression also avoids dependence on explicit crown boundaries, allowing tree numbers to be inferred even when individual crowns are not fully resolved at 3-m pixel resolution. This provides a more direct, spatially explicit baseline for restoration planning and monitoring.
The model analyzed 96 × 96 m image patches from a 20-band composite containing four Planet optical bands, ten Sentinel-2 multispectral bands, and six Sentinel-1 radar polarization bands, all resampled to a common 3-m grid. A hybrid UNet-ResNet50 architecture predicted one tree-count value per patch. In architecture comparisons, UNet-R50 achieved a mean absolute error of 17.32 and an r² of 0.73, outperforming ResNet50, ResNeXt50, SwinT, and ViT. Sensor-ablation tests showed that Sentinel-1 plus Sentinel-2 reached an MAE of 18.42 and r² of 0.77, comparable with the three-sensor composite at 18.8 and 0.77, respectively. Training used about 12 million tree-crown centers from diverse landscapes. The national predictions were then bias-corrected using independent National Forest Inventory measurements, with country- and land-cover-specific bootstrap calibration. Final estimates indicated median densities of 37 trees per hectare in Rwanda and 42 in Tanzania, demonstrating national applicability across contrasting ecosystems. The approach therefore combines scalability with ecological and spatial detail.
The researchers emphasized the framework's practical value, noting that it "can guide the planning and optimization of resource allocation for tree restoration efforts at national scales." They further argue that repeatable satellite-based baselines could help governments track restoration outcomes, identify appropriate planting areas, and better account for trees growing beyond conventional forest boundaries.
Researchers assembled labels from manual interpretation of submeter imagery and semi-automated extraction from airborne and UAV LiDAR canopy-height models. About 12 million crown-center points covered 2,832 sample areas spanning roughly 608,483 hectares. Satellite inputs were temporally harmonized, corrected, normalized, and resampled to 3 m. The network used an 80/20 training-validation split and mean absolute error as its loss function. Predictions were subsequently evaluated and calibrated against independent National Forest Inventory plots from Rwanda and Tanzania. Large trees were used for validation.
The framework could support restoration programs by establishing repeatable national baselines, tracking tree gains and losses, and guiding planting toward ecologically appropriate locations. Future work must improve performance in dense closed-canopy forests, where crown overlap and signal saturation reduce sensitivity, and incorporate understory or smaller woody plants. The authors propose integrating forest inventories and very-high-resolution data, while exploring attention-based sensor fusion, domain adaptation, and transfer learning to extend the approach to new regions, ecosystems, and satellite sensors at larger scales.
###
References
DOI
Original Source URL
https://spj.science.org/doi/10.34133/remotesensing.0970
Funding information
This study received funding from the European Research Council (ERC) under the European Union's Horizon 2020 Research and Innovation Programme (grant agreement no. 101157661 TREEMAP) and the Danish National Research Foundation, Center for Remote Sensing and Deep Learning of Global Tree Resources (TreeSense), DNRF192. LiDAR data from Tanzania were provided by Prof. Terje Gobakken and Prof. Erik Næsset, Norwegian University of Life Sciences. The data were ac-quired as part of the project "Enhancing the measuring, reporting and verification (MRV) of forests in Tanzania through the ap-plication of advanced remote sensing techniques" funded by the Royal Norwegian Embassy in Tanzania as part of the Norwegian International Climate and Forest Initiative. G.V.L. acknowledges the Next Generation EU Mission 4, Component 2 - CUP B53C22002150006 - Project IR0000032 - ITINERIS - Italian Integrated Environmental Research Infrastructures System. The Ghana/Sierra Leone dataset has been collected and preprocessed by the Forest Ecology Lab of Prof. Valentini, DIBAF, Tuscia University, Viterbo, Italy. We also thank Prof. Andrew Davies for providing UAV LiDAR data in the Kruger National Park, South Africa. Additional South African airborne LiDAR data were funded by the Council for Scientific and Industrial Research (CSIR) Strategic Research Panel, the National Research Foundation (South Africa) Global Change Grand Challenge Grant (Grant number 110496), and the University of the Witwatersran.
About J ournal of R emote S ensing
The J ournal of R emote S ensing , an online-only Open Access journal published in association with AIR-CAS, promotes the theory, science, and technology of remote sensing, as well as interdisciplinary research within earth and information science.
Journal of Remote Sensing
Not applicable
A Remote Sensing-Based Solution for National Scale Tree Counting
18-Aug-2026
The authors declare that they have no competing interests.