{"id":1175,"date":"2026-10-28T10:05:05","date_gmt":"2026-10-28T10:05:05","guid":{"rendered":"https:\/\/symposium.earsel.org\/46th-symposium-Turin\/?page_id=1175"},"modified":"2026-11-03T09:44:46","modified_gmt":"2026-11-03T09:44:46","slug":"new-frontiers-in-remote-sensing-of-forests","status":"publish","type":"page","link":"https:\/\/symposium.earsel.org\/46th-symposium-Turin\/?page_id=1175","title":{"rendered":"New frontiers in remote sensing of forests"},"content":{"rendered":"\n
\n

Remote sensing in forest science is rapidly evolving through increasing data availability,
community-driven databases, and advances in artificial intelligence. Emerging initiatives such
as deadtrees.earth, 3Dtrees.earth, and TreeAI offer new opportunities to collect, harmonise,
and share reference data on forest structure, tree species composition, mortality, and related
forest attributes. Such datasets can enable more robust benchmarking and support the
development of automated and transferable forest mapping and monitoring approaches.
Together, these developments are opening new avenues for forest remote sensing that extend
beyond traditional sensor- and region-specific approaches.
At the same time, foundation models, self-supervised representation learning, and other
emerging AI approaches are creating new opportunities to exploit large and heterogeneous
remote sensing datasets. Their potential to transfer across regions, forest types, sensors, and
tasks is particularly promising, while their practical added value, robustness, and data
requirements remain insufficiently explored. Understanding when these approaches provide
genuine benefits over established methods, and how they can be made reliable and
transferable across diverse forest environments, is an emerging research challenge.
This session invites contributions on emerging datasets, collaborative data initiatives, and
novel AI approaches for forest remote sensing. We welcome studies using optical, LiDAR,
radar, hyperspectral, or multimodal data for applications including tree-species identification,
3D forest structure, biomass estimation, forest health and mortality assessment, forest
inventory, and disturbance monitoring. Contributions evaluating foundation models, transfer
learning, self-supervised approaches, multimodal learning, or comparing emerging methods
with established machine-learning techniques are particularly encouraged.
The session aims to connect methodological innovation with open and collaborative research,
promote data and knowledge sharing, and discuss how emerging approaches can contribute
to more transferable, scalable, and reliable forest monitoring and support practical applications
in forest management and planning. It will provide a forum to discuss not only methodological
advances, but also challenges related to reference data quality, benchmarking, reproducibility,
and transferability, as well as the practical usability of emerging remote sensing products and
methods for forest practitioners and decision-makers. We particularly encourage contributions
from early-career researchers and aim to foster scientific exchange around emerging ideas,
datasets, and methodological challenges.<\/p>\n<\/div>\n\n\n\n

\n

Scientific Committee<\/strong><\/h4>\n\n\n\n