New AI system is transforming curlew monitoring - Game and Wildlife Conservation Trust

Conservation work to detect and protect curlew and other ground-nesting birds in Wales have been given a helping hand by new AI monitoring technology used by researchers from the Game & Wildlife Conservation trust and Liverpool John Moores University.

In order to better recognise curlew and their chicks, they trained a new real-time object detection model called YOLOv10 to spot and recognise not only curlew, but also 25 other species, including foxes, deer and rabbits, by combining a pre-existing conservation AI database with nearly 39,000 images from across the UK.

The results of the study, which saw researchers test the system’s ability to monitor curlew and their chicks on 11 sites across Wales during the 2024 nesting season, were presented in a new scientific paper published this spring.

Using the 3G and 4G networks, 1,072 images were sent from AI-enabled cameras through to the technology system, which then processed the curlew footage in real-time, making it accessible through an app on a mobile phone.

They proved that the model was able to filter out blank images triggered by moving vegetation and could also reliably identify curlew chicks, despite their natural camouflage. It correctly identified curlew over 90% of the time and never falsely detected them when they were not there.

To help the AI model handle real-world challenges, the team taught it to recognise animals under different conditions, such as changes in lighting, angles and sizes. This was done through data augmentation, a technique that adjusts images by changing colours, brightness and flipping them to create variety.

In the past, conservation efforts for vulnerable ground-nesting birds like curlews have often been labour intensive and logistically challenging, as the work to process the vast number of images from camera traps is time-consuming, delaying conservation action.

Making use of this new AI assisted technology will significantly reduce conservationists’ workload, allowing them to respond more effectively to protect the sites where breeding curlew have been detected.