Motion-activated cameras, or camera traps, are being increasingly deployed to monitor wildlife populations and inform conservation policy. The amount of data generated by the growing number of camera installations poses a daunting challenge for the researchers who must track and analyze thousands of images. A new data management and sharing platform which also invites participation by citizen scientists has been launched by Google to streamline this task.

The Wildlife Insights online database will allow users to upload camera trap images to Google Cloud, apply Animals in photos are automatically identified using machine learning technology. Source: Wildlife Insightsspecies identification artificial intelligence (AI) models over the images, collaborate with others and develop insights on species population health. Users will be able to ask the system to search for an animal of interest, and all of the images will be publicly available.

The AI model is trained on images from Conservation International's Tropical Ecology and Monitoring Network, Snapshot Serengeti, Caltech Camera Traps, North American Camera Trap Images, WWF and One Tam, and currently includes 614 species from around the world. The system catches 78.7% of blank images with an error rate of less than 2%, allowing program users to minimize the amount of human involvement in the process of sifting through millions of images looking for wildlife.

The beta version of Wildlife Insights has been developed by a partnership that also includes Conservation International, the Smithsonian’s National Zoo and Conservation Biology Institute, Wildlife Conservation Society, North Carolina Museum of Natural Sciences, World Wide Fund for Nature, Zoological Society of London and Yale University’s Map of Life.