A first-of-its-kind study combines satellite imagery and AI to monitor the endangered species.

Scientists from the University of Bath, University of Oxford, and the University of Twente in the Netherlands have demonstrated the successful use of satellite cameras with deep learning algorithms to monitor the location and number of African elephants. (Science Focus)

Tracking the diminishing numbers of African elephants has become critical for animal conservationists. Loss of habitat and unbridled poaching has decimated the numbers of this magnificent animal, now classified as an endangered species. According to one estimate, there are just 50,000 African elephants now left in the wild.

How it works

The new system is a huge improvement over the current method of counting elephants manually from low-flying aircraft.

Created by Dr. Olga Isupova, a computer scientist at the University of Bath, the automated artificial intelligence system analyses high-resolution images of elephants taken by the commercial Worldview-3 observation satellite. It does not matter that the images were captured whilst the animals were on the move in their natural habitat of forests and grasslands.

The system could identify the African elephants as well as the human method. Interestingly, similar techniques have been used for tracking marine species such as whales. However, this is the first time that it has been applied to terrestrial wildlife.

Elephants are harder to track than whales

“This type of work has been done before with whales, but of course the ocean is all blue, so counting is a lot less challenging. As you can imagine, a heterogeneous landscape makes it much hard to identify animals,” said Dr. Isupova.

The study could become more accurate in the future because satellite imagery resolution increases every couple of years.

Soon, therefore, conservationists would be able to see the wildlife in far more detail. The technology could then extend to many more different species.

“Accurate monitoring is essential if we are to save the species,” says Dr. Isupova. “We need to know where the animals are and how many there are.”

Related Story:  Machine Learning To Identify Whale Sounds, Warn Ships