African Penguins are critically endangered, with fewer than 20,000 breeding pairs remaining compared to historical estimates exceeding one million. Effective conservation requires continuous population monitoring, but traditional methods are labour-intensive and cannot scale to meet current needs. This project develops practical artificial intelligence tools that enable automated monitoring of penguin populations using camera imagery. The core innovation addresses a fundamental problem: training AI models typically requires thousands of manually annotated images, which is prohibitively expensive for most conservation programmes. My research demonstrates that intelligent data selection combined with semi-automated annotation can reduce this requirement by over 40% while actually improving model accuracy. I will refine these methods, validate them across multiple monitoring sites, and deploy working systems with South African National Parks. This work establishes foundations for broader application to other endangered species across Africa.
2027