Canelas, J., Clementino, L., Cid, A., Castro, J., Machado, I., Vieira, S. (2025). Automated cetacean detection in UAV imagery using AI models: a case study on Delphinid species. Int J Data Sci Anal.
https://doi.org/10.1007/s41060-024-00704-9
AIMM’s first open access paper of 2025!
This study looks at how to improve the way we detect marine mammals like dolphins, which is important for understanding their abundance and helping with conservation. Traditional methods are time-consuming and not always accurate, so the researchers tested advanced computer models to automate this process using drone images and videos. They trained these models using examples of dolphins from online videos and tested them in real-world conditions.
The image-based approach worked well, with the best model achieving high accuracy (around 84%) in detecting dolphins in drone photos. However, video-based methods had more challenges, with lower accuracy (68%) because the models struggled to adapt to new environments.
This research demonstrates a promising way to make monitoring dolphins easier and more effective, which could lead to better conservation strategies.
To find out what the discoveries showed, read the article here.