Finnish Team Develops New Remote-Sensing Method to Track Keystone Forest Species at Scale

(Image: Samuli Junttila)

(IN BRIEF) Scientists from the University of Helsinki and the University of Eastern Finland have developed AI-enhanced methods to identify aspens and standing dead trees using open aerial imagery, offering a fast and low-cost approach for biodiversity monitoring. Aspens host more than a thousand species, yet accurate mapping has previously been challenging to perform at scale. The new model recognises aspens with high accuracy year-round and performs especially well for large mature trees. A second study introduced a method for detecting standing deadwood — an essential habitat for many species — using neural network and image-processing techniques. Both studies showcase a major step forward for remote ecosystem monitoring and could support forest management, conservation planning, and future biodiversity research.

(PRESS RELEASE) HELSINKI, 10-Dec-2025 — /EuropaWire/ — Researchers at the University of Helsinki and the University of Eastern Finland have developed new AI-driven methods that allow aspen trees and standing dead trees to be identified automatically from open aerial imagery — a breakthrough that could transform biodiversity monitoring across Finnish forests. Despite making up only a small fraction of Finland’s tree population, aspens (Populus tremula) are known as a keystone species supporting more than a thousand organisms ranging from insects and fungi to birds and mammals. Their nutrient-rich bark and fast-decomposing litter create microhabitats where many forest species thrive, yet mapping them has historically required time-consuming and costly field surveys.

The team led by Associate Professor Samuli Junttila has demonstrated for the first time that aspens can be reliably distinguished at scale using aerial photography and neural network-based image analysis. The technique performs consistently across seasons — even distinguishing aspens with and without leaves — and proved especially effective in detecting tall, mature trees that play a critical role in forest ecosystems. The research, published in Remote Sensing Applications: Society and Environment in October 2025, was led by doctoral researcher Anwarul Chowdhury, who notes the model’s high reliability and practical value for forestry management and conservation work across Finland.

Junttila’s Global Ecosystem Health Observatory research group specialises in combining remote sensing and AI to track forest health. Their tools have previously been used to monitor bark beetle damage and assess tree mortality. Building on this work, a second study published in November 2025 by researchers Anis Rahman, Einari Heinaro, and Mete Ahishali introduced a technique for identifying standing dead trees from aerial data — another indicator of biodiversity. Deadwood is vital for species conservation, yet detecting it remotely is difficult due to canopy cover. The team achieved improved accuracy by using machine-learning models with hybrid self-attention layers and adaptive filtering to analyse aerial images.

Together, these innovations offer scalable monitoring tools capable of mapping forest habitats using freely available national datasets. The researchers plan to further improve juvenile aspen detection, potentially by integrating laser scanning data with aerial imagery to enhance classification precision.

Original articles

The article ‘Mapping large European aspens (Populus tremula L.) using national aerial imagery and a U-Net convolutional neural network’Opens in a new tab was published in October 2025 in the Remote Sensing Applications: Society and Environment journal. https://doi.org/10.1016/j.jag.2025.104851Opens in a new tab

The article ‘Dual-task learning for dead tree detection and segmentation with hybrid self-attention U-Nets in aerial imagery’Opens in a new tab was published in November 2025 in the International Journal of Applied Earth Observation and Geoinformation. https://doi.org/10.1016/j.rsase.2025.101755

Media contact:

Samuli Junttila
Associate Professor
Department of Forest Sciences
samuli.junttila@helsinki.fi

Anwarul Chowdhury
Technical Assistant
Department of Forest Sciences
anwarul.chowdhury@helsinki.fi

Anis Rahman
University Researcher
Department of Forest Sciences
anis.rahman@helsinki.fi

SOURCE: University of Helsinki