How AI-Powered Drones Are Helping Conservationists Save Animals

The EPFL research groups have created systems that utilise drones outfitted with AI-powered computer vision software for tracking and cataloguing wildlife populations in honor of UN World Wildlife Day. This project aims to change wildlife protection by offering timely, useful information.

World Wildlife Day, which was established by the UN ten years ago, honors the approval of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) on March 3 of that year. To commemorate this milestone, we choose to highlight several EPFL research teams that are creating tools for identifying, tracking, and conserving wildlife.

Keeping an Eye on the Skies to Follow the Fauna

Drones, which buzz above the ground, are already a common way of monitoring and categorizing wildlife populations remotely. An example of this is the recent launch of the most recent in a series of projects by Devis Tuia and his team to improve the AI-powered computer vision software that automatically pulls data from drone photographs in the Kuzikus Game Reserve in Namibia.

Prof. Tuia, who is in charge of the Environmental Computational Science and Earth Observation Laboratory at EPFL, played a key role in bringing together researchers from academic institutions, private companies, and non-profit organizations who are involved in computer vision, drone technology, and conservation ecology. Together, they founded the WildDrone network, a global project whose goal is to employ drone technology to completely transform animal protection. Ten PhD projects are included in the effort, with two of them receiving funding from UK Research and Innovation (UKRI) and one from the Swiss State Secretariat for Education, Research, and Innovation (SERI).

Researchers in Namibia’s Kuzikus Game Reserve utilize AI-based tools to estimate animal populations.

According to Tuia, drones have fundamentally altered the way wildlife is monitored. In addition to allowing you to travel farther than, say, a helicopter, they are also less expensive, safer, and easier to scale. Drones with AI-powered computer vision can be used to deploy when necessary and provide actionable insights in close to real-time, in addition to counting animal populations every few months.

There is still room for development, though. According to Tuia, “Now, we can develop AI models that work well in one environment, such as a wildlife reserve, but not so well in another or even the same one at a different time of year. This issue is being addressed by the research his research group is conducting as part of the WildDrone initiative.

The Kuzikus project is just one of many wildlife conservation initiatives being carried out by Tuia’s research team to estimate animal numbers using AI-based technologies. For instance, they used drones and their AI technology to track the population of seabirds that inhabit coastal areas, such as the African Royal Tern. Also, their software is revealing information on local wildlife species interactions in the Swiss National Park.

According to Tuia, “our dream” is to be able to observe animals without endangering or upsetting them and to provide rangers with the real-time information they require to carry out their duty of guarding wildlife. We may be able to go much further in a few years and provide knowledge that will help decision-makers create better policies and resolve conflict.

To determine why Red Sea corals are so resistant to pollution and global warming, several EPFL laboratories are working. Banc-Prandi, Guilhem

Protecting coral reefs

Coral reefs have already been destroyed in half of the world’s coral reefs, taking with them not just their eye-catching colors but also the ecosystems that had developed around them. Everything is being impacted by the decline of coral reefs, from single-cell algae to coastal fishing populations.

Several EPFL labs are researching solutions to protect coral reefs using AI technologies to address this issue. Using aerial photographs captured by drones or satellites, researchers from EPFL’s Artificial Intelligence Lab (AI Lab) have created an AI system that can identify coral reefs that have bleached.

The AI system uses deep learning algorithms to recognize distinctive traits, like hue and texture changes brought on by bleaching occurrences, in aerial photographs captured by drones or satellites and identify bleached corals.

Professor Dario Floreano from EPFL’s Artificial Intelligence Lab commented, “This method could prove invaluable for researchers trying to monitor coral reef health” (AI Lab). They wouldn’t need to physically travel to every reef spot if they used it, which would save them time and resources.

Source: https://innovationorigins.com/en/laio/ai-powered-drones-revolutionize-wildlife-conservation/