A 17-year-old creates an ultra-efficient and cost-effective anti-poaching device to protect elephants

A 17-year-old creates an ultra-efficient and cost-effective anti-poaching device to protect elephants

A 17-year-old creates an ultra-efficient and cost-effective anti-poaching device to protect elephants

⇧ [VIDÉO] You may also like this partner content (after announcement)

An iPhone, a drone and a modest thermal imaging camera… It didn’t take longer for 17-year-old Anika Puri to create an anti-poaching system four times more effective than those currently used to protect elephants.

The idea came to the girl four years ago, explains al Smithsonian magazine. Visiting India with her family, she was surprised to see so many ivory items still on sale in the markets, even though the elephant hunting ban had been in place for many years. Indeed, elephant populations continue to decline around the world. Between 2002 and 2011, the African forest elephant population, for example, decreased by 62%. In 2013 it represented only 10% of its potential, despite the protection allegedly in place since the 1990s.

Monitoring systems are already in place. However, as he discovers, these are often visual recognition based surveillance systems. They are too ineffective, as well as being extremely expensive. In fact, the type of thermal imaging cameras used can cost up to nearly $ 10,000, says the foundation’s website that awarded it.

Anuka Puri then decides to roll up her sleeves. She is passionate about the possibilities offered by the field of artificial intelligence, she uses her knowledge on the subject. In particular, she uses machine learning. It is a technology that allows a program to “learn” by feeding on data, in order to then be able, for example, to analyze situations on its own.

She has managed to create a system that uses fairly simple tools: an iPhone and a relatively inexpensive thermal imaging camera, all connected to a drone. According to her, the system could cost around € 300. All the interest of the project therefore lies not in the physical structure itself, but much more in the software that she has developed with care. Instead of relying on a visual analysis of the filmed species, she focused instead on the movements of elephants and humans. ” I realized that we could take advantage of this disparity between these two movement patterns to increase accuracy in detecting potential poachers. “He explains to Smithsonian Magazine.

Analyze the movements of elephants and humans

During a conference on artificial intelligence, he meets Elizabeth Bondi-Kelly, a computer scientist at Harvard. The latter is working on a wildlife conservation project using drones and machine learning. She then became his mentor for this project and allowed him to use a database collected in the field: the “Benchmarking IR Dataset for Surveillance with Aerial Intelligence” (BIRDSAI).

Using a machine learning algorithm, he created a model that can determine whether the movements are made by elephants or humans, based on speed, group size and many other things. He used 372 datasets, including 300 elephant movements and 72 human movements. The remaining 144 were used to test her model with data she had never seen before. Her software then showed 90% accuracy in detecting humans. A method therefore four times more effective, she says, than those already in place.

His invention, called ElSa, for Elephant Savior, earned him an award, particularly for his clear explanation of the project: the Peggy Scripps Award for Science Communication. He also won a prestigious award in the earth and environmental sciences category of the competition. The girl is not going to stop on such a good road. In the future, she wants to continue to test the effectiveness of her project and implement it on the ground in Africa. She also aims to extend it to other endangered species, such as rhinos. She also plans: the optimization of the flight paths of drones, which have a limited autonomy.

Leave a Reply

Your email address will not be published.