Capturing the power of AI for wildlife conservation.
More than a quarter of all known species are threatened by extinction. To protect them we need to know where they are, how large their populations are, and what’s happening in their local environment. Ecologists gather this information through a combination of fieldwork and remote sensing aided by technology.
Camera traps are one tool that’s used across the globe to capture images of species in their native habitats. Millions upon millions of camera trap images have been taken, but sorting and analyzing them is time-consuming, and the data they produce can exist in siloes. Wildlife Insights uses machine learning to automatically identify species captured in camera trap images and provides tools that analyze wildlife trends. With these tools, researchers are able to make better decisions and share their findings more easily.
The founding and core members behind Wildlife Insights includes Conservation International, Wildlife Conservation Society, Google, North Carolina Museum of Natural Sciences, Smithsonian Institution, Zoological Society of London, WWF, and Map of Life. Our role as a project partner was to build the platform from scratch, ensuring it would remain stable and scalable as it grows.
Early users of Wildlife Insights estimate that they have reduced the amount of time they spend processing and reviewing images by 80%.
Wildlife Insights needs a solid foundation upon which the platform can grow. To achieve this we custom-built the backend using NestJs for the main API and deployed it to a Kubernetes cluster. All of it runs on Google Cloud.
In addition to Google Cloud we’re using Airflow to run the data analysis pipelines, and Google’s AI Platform for the computer vision. Many of the processes that drive Wildlife Insights are delegated to cloud functions, which means the system will remain performant even as the already massive dataset grows.
Faster species identification.
Wildlife Insights is the largest shared camera trap data set in the world.
Common classes of species can be identified with an accuracy of between 80% and 98.6%.
Explore the data on an interactive map, analyze wildlife trends, or make customized reports.
What are people saying about it?