In zoology, cutting-edge technologies can be rare, especially when it comes to keeping track of animals such as elephants across vast lands. However, thanks to deep learning and satellite imagery from Maxar Technologies, an algorithm can pick out elephants from a satellite image.
University of Oxford researchers used Maxar’s WorldView-3 satellite imagery and deep learning (TensorFlow API, Google Brain) to detect and track elephants in South Africa with comparable accuracy to that of human detection. Maxar’s WorldView-3 satellites can collect up to 5,000 sq km imagery in one pass within minutes. This drastically reduces errors and potential double counting while being unobtrusive to both animals and humans.
Currently, animal conservationists use aircraft to survey lands and identity elephants, but the new technology of satellite imagery and AI can expand the monitoring across international borders, where it is costly and difficult to get permission for aircraft surveys. Maxar Technologies’ satellite imagery coupled with University of Oxford’s deep learning can help save money, time, decrease poaching, track and protect the world’s elephants.