At this year’s Alternet Conference on Biodiversity and Ecosystem Services (13–16 May 2025, Aveiro, Portugal), OneSTOP partner Toke Thomas Høye (Aarhus University) delivered a keynote titled “Automated species monitoring from the Arctic to the tropics.” His talk focused on the application of computer vision and deep learning in camera-based biodiversity monitoring across various ecosystems.

Toke presented how modern camera systems can be used to continuously and non-invasively observe insects, plants, and other organisms across diurnal and seasonal cycles. He outlined how deep learning models can estimate key ecological metrics such as abundance, biomass, and diversity from image data, providing a scalable approach to tracking biodiversity change.

Drawing from large-scale monitoring programmes across national and continental levels, Toke illustrated the potential of automated observation systems to generate rich, multidimensional datasets. He also illustrated how projects such as MAMBO and Biodiversa+ are using the AMI system, which will also be tested across OneSTOP’s Living Labs.

Toke also discussed ongoing technical challenges and outlined future research directions needed to support wider implementation of sensor-based monitoring, particularly for plants and insects.