CHMv2: Improvements in Global Canopy Height Mapping using DINOv3

2026-03-06

AI Summary

This study addresses the global scarcity of high-accuracy forest canopy height data, where existing remote sensing products suffer from systematic underestimation in tall forests and fail to preserve fine-scale structural features such as canopy edges and gaps. To overcome these limitations, the authors propose CHMv2, a novel canopy height mapping approach that leverages the DINOv3 vision foundation model to fuse high-resolution optical satellite imagery with airborne laser scanning (ALS) data. The method incorporates a geographically diverse training strategy, an automated data cleaning and co-registration pipeline, and a tailored loss function combined with a height-distribution-aware sampling mechanism. Evaluated at 1-meter resolution globally, CHMv2 achieves substantially improved accuracy and structural fidelity across diverse forest biomes, consistently outperforming current products and effectively mitigating the persistent underestimation of canopy heights in tall forests.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Ensemble Methods
• Data Mining & Knowledge Management: Mining of Visual, Multimedia & Multimodal Data

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Accurate canopy height information is essential for quantifying forest carbon, monitoring restoration and degradation, and assessing habitat structure, yet high-fidelity measurements from airborne laser scanning (ALS) remain unevenly available globally. Here we present CHMv2, a global, meter-resolution canopy height map derived from high-resolution optical satellite imagery using a depth-estimation model built on DINOv3 and trained against ALS canopy height models. Compared to existing products, CHMv2 substantially improves accuracy, reduces bias in tall forests, and better preserves fine-scale structure such as canopy edges and gaps. These gains are enabled by a large expansion of geographically diverse training data, automated data curation and registration, and a loss formulation and data sampling strategy tailored to canopy height distributions. We validate CHMv2 against independent ALS test sets and against tens of millions of GEDI and ICESat-2 observations, demonstrating consistent performance across major forest biomes.

Problem

• canopy height
• global mapping
• forest structure
• high-resolution
• accuracy

Innovation

• canopy height mapping
• DINOv3
• depth estimation
• airborne laser scanning
• global forest monitoring