FLORA: A deep learning approach to predict forest attributes from heterogeneous LiDAR data

2026-06-30

AI Summary

This study addresses the degraded generalization performance in forest attribute prediction caused by sensor heterogeneity, flight parameters, seasonal variations, and scanning angle differences in multi-source airborne LiDAR data. To overcome these challenges, we propose FLORA, a novel framework built upon an octree-based neural network backbone that, for the first time, enables a single model to jointly process both leaf-on and leaf-off LiDAR point clouds. FLORA integrates ecological and spatiotemporal auxiliary variables and employs a late-fusion gated mechanism to enhance cross-seasonal robustness. Evaluated on 32,052 field plots across mainland France, FLORA significantly outperforms existing methods at the national scale, achieving a relative RMSE of 12.3% (R² = 0.88) for dominant height and 39% (R² = 0.74) for total timber volume, thereby establishing a reliable baseline for large-scale, high-accuracy forest monitoring.

Abstract

Forest attributes are essential for national-scale resource monitoring. Airborne LiDAR metrics are among the auxiliary variables most strongly correlated with forest attributes used in National Forest Inventory (NFI) estimates. However, producing wall-to-wall predictions remains challenging when LiDAR data are acquired under heterogeneous conditions. As national LiDAR programs expand across Europe, variability in sensors, flight parameters, seasons, and scan angles limits the robustness of existing models, which are often calibrated for local conditions. We present FLORA (Forest LiDAR Octree Regression with Auxiliary Data), a deep learning framework that predicts six forest attributes: dominant height, total volume, deciduous volume, coniferous volume, basal area, and stem density from heterogeneous LiDAR point clouds. FLORA combines an octree-based backbone with ecological and spatiotemporal auxiliary variables through a late-fusion gating mechanism. Models are trained and evaluated on 32,052 National Forest Inventory plots across mainland France using data from the French LiDAR HD program. A single model trained on both leaf-on and leaf-off acquisitions outperforms season-specific models and improves cross-season robustness. Auxiliary variables provide modest overall gains but contribute more strongly to species-specific volume prediction. FLORA achieves an rRMSE of about 12.3% (R2 = 0.88) for dominant height and 39% (R2 = 0.74) for total volume, providing a robust baseline for large-scale forest attribute estimation from heterogeneous national LiDAR programs.

Problem

• heterogeneous LiDAR data
• forest attributes
• National Forest Inventory
• cross-season robustness
• wall-to-wall prediction

Innovation

• octree-based deep learning
• heterogeneous LiDAR
• late-fusion gating
• cross-season robustness
• forest attribute prediction