When AI Counts Penguins: High-Tech in Antarctica
The rapid development of artificial intelligence (AI) is opening up new possibilities for scientific research, particularly in the field of ecology. An outstanding example is an innovative project by the Chilean Antarctic Institute (INACH), which focuses on the automated monitoring of penguin populations. Under the leadership of marine biologist Magdalena Márquez Díaz, AI is being used to analyze the life cycle of chinstrap penguins (Pygoscelis antarcticus) more efficiently and precisely.
The project is part of the Marine Protected Areas (MPA) program and aims to better understand and protect Antarctic ecosystems. With the help of camera traps and modern image processing systems, continuous data collection is possible without researchers needing to be permanently on site. This method represents a significant advancement over traditional, time- and labor-intensive observation techniques.
At the core of the project is an AI algorithm based on the neural network YOLO (You Only Look Once). This computer vision system was trained using annotated image data to distinguish between adult penguins and chicks. As a result, populations can be counted automatically, and key biological parameters such as reproductive success, arrival and departure times, and duration of stay in colonies can be determined.
The AI analyzes thousands of images captured by camera traps each year. While manual evaluation used to take weeks, data processing is now completed within just a few hours. This not only increases efficiency but also improves the accuracy of scientific results.
Since 2022, several camera traps have been installed along the Antarctic Peninsula, including on Kopaitic Island and Nelson Island. These devices operate according to the standards of the Commission for the Conservation of Antarctic Marine Living Resources as part of the Ecosystem Monitoring Program (CEMP). Once a year, the storage modules are replaced, and the data is transported to Punta Arenas for analysis.
Despite the positive results, the project faces technical challenges. Detecting chicks is particularly difficult, as they are often obscured by adult birds or are hard to distinguish from the background due to their camouflage. Nevertheless, the system is already delivering promising results that are continuously being improved.
The iterative nature of AI systems allows for ongoing optimization. In the future, additional features are planned, such as recognizing breeding behavior, monitoring survival rates, and analyzing interactions with other species like skuas. The identification of different penguin species using new AI methods is also planned.
This project highlights the enormous potential of artificial intelligence in biodiversity research. By using automated systems, large volumes of data can be processed efficiently and new scientific insights can be gained. At the same time, the need for complex field missions is reduced, saving costs and minimizing the impact on sensitive ecosystems.
Conclusion
The use of AI in monitoring chinstrap penguins represents a significant advancement in Antarctic research. The INACH project serves as an excellent example of how modern technologies can help address complex ecological questions. With further improvements and expansions, this method is expected to play an even more important role in protecting and understanding Antarctic biodiversity in the future.
Rosamaria Kubny, PolarJournal
