# Assessing the Effectiveness of Deep Embeddings for Tree Species Classification in the Dutch Forest Inventory

arXiv: https://arxiv.org/abs/2508.18829 · HTML: https://mlsanity.com/paper/2508.18829/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 68 references that are in the map.
- First posted on arXiv 26 Aug 2025.

## Builds on (most cited first, 18 of 18)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Foundation Models for Generalist Geospatial Artificial Intelligence](https://mlsanity.com/paper/2310.18660/) (2023)
- [Prithvi-EO-2.0: A Versatile Multi-Temporal Foundation Model for Earth Observation Applications](https://mlsanity.com/paper/2412.02732/) (2024)
- [Self-supervised Learning in Remote Sensing: A Review](https://mlsanity.com/paper/2206.13188/) (2022)
- [Lightweight, Pre-trained Transformers for Remote Sensing Timeseries](https://mlsanity.com/paper/2304.14065/) (2023)
- [A high-resolution canopy height model of the Earth](https://mlsanity.com/paper/2204.08322/) (2022)
- [Temporal Convolutional Neural Network for the Classification of Satellite Image Time Series](https://mlsanity.com/paper/1811.10166/) (2018)
- [Self-attention for raw optical Satellite Time Series Classification](https://mlsanity.com/paper/1910.10536/) (2019)
- [Deep Learning in Multimodal Remote Sensing Data Fusion: A Comprehensive Review](https://mlsanity.com/paper/2205.01380/) (2022)
- [Satellite Image Time Series Classification with Pixel-Set Encoders and Temporal Self-Attention](https://mlsanity.com/paper/1911.07757/) (2019)
- [Multi-Temporal Land Cover Classification with Sequential Recurrent Encoders](https://mlsanity.com/paper/1802.02080/) (2018)
- [Early Stopping without a Validation Set](https://mlsanity.com/paper/1703.09580/) (2017)
- [TESSERA: Temporal Embeddings of Surface Spectra for Earth Representation and Analysis](https://mlsanity.com/paper/2506.20380/) (2025)
- [DeepSat - A Learning framework for Satellite Imagery](https://mlsanity.com/paper/1509.03602/) (2015)
- [An Empirical Analysis of Feature Engineering for Predictive Modeling](https://mlsanity.com/paper/1701.07852/) (2017)
- [High Resolution Tree Height Mapping of the Amazon Forest using Planet NICFI Images and LiDAR-Informed U-Net Model](https://mlsanity.com/paper/2501.10600/) (2025)
- [Tree species classification at the pixel-level using deep learning and multispectral time series in an imbalanced context](https://mlsanity.com/paper/2408.08887/) (2024)

## Built on by (most cited first, 1 of 1)

- [Earth Embeddings](https://mlsanity.com/paper/2608.03410/) (2026)

## Abstract (arXiv)

National Forest Inventory (NFI) serves as the primary source of forest information, however, maintaining these inventories requires labor-intensive on-site campaigns by forestry experts to identify and document tree species. Embeddings from deep pre-trained remote sensing models offer new opportunities to update NFIs more frequently and at larger scales. This work systematically investigates how deep embeddings improve tree species classification accuracy in the Netherlands with few annotated data. We evaluate this question on three embedding models: Presto, Alpha Earth, and TESSERA, using three tree species datasets of varying difficulty. Data-wise, we compare the available embeddings from Alpha Earth and TESSERA with dynamically calculated embeddings from a pre-trained Presto model, for which we extracted time series from Sentinel-1 , Sentinel-2, and weather data, along with elevation data downloaded from Google Earth Engine. Our results demonstrate that publicly available remote sensing time series deep embeddings outperform the current state-of-the-art hand crafted features in NFI species classification in the Netherlands, yielding performance gains of roughly 7 to 9 percentage points in overall accuracy and up to 17 points in macro-F1 on the NFI datasets at the reference plot level. This indicates that classic hand defined features are too simple for this task and highlights the potential of using deep embeddings for data-limited applications such as NFI classification. Country-scale maps built from the pre-computed embeddings reach accuracy consistent with the plot-level results and with an independent comparison against the NFI. By leveraging openly available satellite data and deep embeddings from pre-trained models, this approach consistently improves classification accuracy compared to traditional methods and can effectively complement existing forest inventory processes.