Artificial intelligence is not only transforming industry and medicine; it is also becoming a key ally in the conservation of endangered species. One example of this is the WADAS project (Wild Animal Detection and Alert System), an initiative led by the Salviamo l’Orso association, the Abruzzo, Lazio and Molise National Park (PNALM), the University of L’Aquila, and Intel, which aims to protect the rare Marsican brown bear, a subspecies of the European brown bear that currently has just 50 individuals left in the wild.
This intelligent detection and alert system combines Reolink Go Ranger PT cameras with Intel® Core™ Ultra 200V processors and the OpenVINO™ AI inference tool. The result is an efficient, portable solution that requires no GPU or cloud infrastructure and is capable of identifying the presence of wildlife in real-time in high-risk areas or close to urban centers.
Low-Power AI for Natural Environments
The WADAS system leverages the potential of the new Intel Core Ultra 200V processors, which integrate a neural processing unit (NPU) and Intel Arc™ graphics. This combination efficiently divides the workload: the NPU detects elements such as animals, people, or vehicles, while the GPU classifies species like bears, deer, or wolves.
Thanks to this local AI architecture, the system can function even in rural areas without continuous connectivity, alerting park rangers or authorities in real time. It also includes the activation of IoT actuators, such as traffic lights on roads or smart feeders, which can mitigate conflicts with wildlife and reduce accidents.
“The Abruzzo National Park represents one of the oldest protected natural spaces in Italy, where endangered species coexist with local communities in a delicate balance. AI can safely reinforce this coexistence,” explained Stefano Dell’Osa, functional safety software architect at Intel and a volunteer for the project.
Accessible and Open Source Technology for a Vital Cause
The development of WADAS has been designed as an open-source project, intended to run on personal computers, laptops, or compact devices with Intel processors, without requiring expensive servers or external graphics cards. This facilitates its replicability in other regions of the world with similar issues.
Initial field tests conducted in 2025 have demonstrated 97.4% accuracy in wildlife detection and classification, confirming the system’s potential as a robust tool for biodiversity protection.
“We are proud to have propelled WADAS, a project born from the perseverance of our volunteers, which addresses one of the main threats to the Marsican bear: road accidents,” explained Stefano Orlandini, president of Salviamo l’Orso.
Reolink Cameras Adapted for Wildlife
The optical hardware is another pillar of the project. The Reolink Go Ranger PT cameras, with non-intrusive night vision and 4K recording, offer precise observation without disrupting animal behavior. They use invisible infrared at 940 nm and an f1.6 aperture to capture footage in complete darkness, with panning and tilting movements to cover large areas.
An Initiative with Global Potential
The accumulated experience of PNALM and local associations is key to adapting technology to the real needs of the environment. Complementary developments, such as automated feeders and reactive signage, are already being designed to reduce risks in areas where wildlife approaches urban zones.
In the words of Luciano Sammarone, director of PNALM, “technology, when used with sensitivity and knowledge, becomes an essential tool for conservation and biodiversity management.”
WADAS not only aims to protect the Marsican brown bear but also to lay the groundwork for a replicable strategy in other parks and with other species. It is a concrete example of how AI, far from being a threat, can be an ally in protecting wildlife.
References: Intel News, WADAS Announcement (Salviamo L’Orso) | Wild Animals Detection and Alert System (WADAS) and OpenVINO: AI-Powered Wildlife Protection.