Keeping an accurate count of at-risk species is an important task. Counting elephants in Africa can be difficult, though, with aircraft surveys taking lots of time, costing a lot of money, and involving border-related challenges. Not to mention the fact that it’s not always accurate, thanks to human error. Scientists at Oxford University are using another counting method: satellites.
A team from the University of Oxford Wildlife Conservation Research Unit and Machine Learning Research Group has been combining the highest resolution satellite capabilities with deep learning to monitor these iconic animals from space. The satellites can collect more than 5,000 km² of images in just one pass in a few minutes, which does away with the risk of double counting and allows for a quicker repeat survey.
Being able to pinpoint population numbers like this will ensure that conservation funding is distributed appropriately and that population trends are accurately understood.
A news release from the university explains, “The population of African elephants has plummeted over the last century due to poaching, retaliatory killing from crop raiding and habitat fragmentation. To conserve them requires knowledge of where they are, and how many there are: accurate monitoring is vital.”
The Oxford team, in collaboration with Dr. Olga Isupova from the University of Bath and Dr. Tiejun Wang from the University of Twente, created a customized training dataset of more than 1,000 elephants in South Africa. This was fed into an algorithm known as a Convolutional Neural Network, which allowed the team to compare the deep learning model’s elephant detection to that of humans. The results found the two were comparable. The model was also able to find elephants in places outside the training data site and identify calves.
Dr. Isupova says, “We just present examples to the algorithm and tell it, ‘This is an elephant, this is not an elephant.’ By doing this, we can train the machine to recognize small details that we wouldn’t be able to pick up with the naked eye.”
One of the reasons satellite monitoring can be problematic is because of the need to process large amounts of images. If it’s not automated, it could take months. However, with this automation, that can be cut down to a few hours. Oxford says machines are also less apt to make an error, and issues with false negatives or false positives can be fixed by improving models. The method has the added benefit of not disrupting species and not putting humans at risk while data is collected, as well.
Dr. Isla Duporge, an Oxford conservation scientist, says, “In zoology, technology can move quite slowly. So being able to use the cutting-edge techniques for animal conservation is just really nice.”
Article continues below
Our Featured Programs
See how we’re making a difference for People, Pets, and the Planet and how you can get involved!
Duporge added that some conservation agencies are already interested in using this technology to replace aircraft counts.
The World Wildlife Fund says that elephant populations have sharply decreased since the beginning of last century. There are some stable and growing populations, but the species is still threatened due to the ivory trade, conflicts with humans, and habitat destruction.Whizzco