Using Computer Vision to Save Sea Lions

Researchers are employing computer vision technology to aid the conservation of the endangered Steller sea lion, whose population in the western Aleutian Islands has declined dramatically by 94% over the past three decades. Despite the puzzling decline in this region, sea lion populations elsewhere are increasing, prompting efforts to study their behaviors, such as movement, survival, and birth rates, to identify potential environmental factors affecting them. To facilitate this research, NOAA biologists mark a small portion of sea lions and use remote cameras to capture images, but the manual process of sifting through these images is labor-intensive. Steller Sight, a tool developed with Roboflow's support, automates the filtering of these images, accurately identifying whether sea lions are present with a high accuracy rate of 97% and significantly reducing the time researchers spend on image review. This technology not only speeds up data collection but also holds potential for broader applications in wildlife conservation, including monitoring killer whale predation, which could impact sea lion populations.