Deep Pre-trained Time Series Features for Tree Species Classification in the Dutch Forest Inventory

2025-08-26

AI Summary

Tree species classification in the Dutch National Forest Inventory is hindered by scarce labeled training data and limited discriminative power of handcrafted features. Method: We propose a deep pretraining-based feature transfer approach leveraging multi-source time-series remote sensing data—Sentinel-1/2, ERA5, and SRTM—uniformly extracted via Google Earth Engine. A publicly available remote sensing time-series pretrained model is fine-tuned to generate highly discriminative deep temporal features. Contribution/Results: This work presents the first systematic validation of pretrained time-series foundation models for few-shot tree species classification, demonstrating superior generalization and effectiveness under data scarcity. Experiments show up to a 10 percentage-point improvement in classification accuracy over state-of-the-art methods, substantially alleviating the data bottleneck. The proposed framework establishes a scalable deep learning paradigm for forest resource inventory and monitoring.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Remote Sensing / Geospatial AI
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

National Forest Inventory (NFI)s serve as the primary source of forest information, providing crucial tree species distribution data. However, maintaining these inventories requires labor-intensive on-site campaigns. Remote sensing approaches, particularly when combined with machine learning, offer opportunities to update NFIs more frequently and at larger scales. While the use of Satellite Image Time Series has proven effective for distinguishing tree species through seasonal canopy reflectance patterns, current approaches rely primarily on Random Forest classifiers with hand-designed features and phenology-based metrics. Using deep features from an available pre-trained remote sensing foundation models offers a complementary strategy. These pre-trained models leverage unannotated global data and are meant to used for general-purpose applications and can then be efficiently fine-tuned with smaller labeled datasets for specific classification tasks. This work systematically investigates how deep features improve tree species classification accuracy in the Netherlands with few annotated data. Data-wise, we extracted time-series data from Sentinel-1, Sentinel-2 and ERA5 satellites data and SRTM data using Google Earth Engine. Our results demonstrate that fine-tuning a publicly available remote sensing time series foundation model outperforms the current state-of-the-art in NFI classification in the Netherlands by a large margin of up to 10% across all datasets. This demonstrates that classic hand-defined harmonic features are too simple for this task and highlights the potential of using deep AI features for data-limited application like NFI classification. By leveraging openly available satellite data and pre-trained models, this approach significantly improves classification accuracy compared to traditional methods and can effectively complement existing forest inventory processes.

Problem

• Classifying tree species using satellite time series data
• Improving accuracy with deep pre-trained remote sensing features
• Reducing reliance on labor-intensive forest inventory methods

Innovation

• Pre-trained deep features from remote sensing foundation models
• Fine-tuning with limited annotated satellite time-series data
• Leveraging multi-source satellite data for improved classification