Innopolis Trains AI to Prevent Collisions With Wild Animals
Innopolis-based company Invian is developing a neural network designed to detect wild animals in urban areas and on highways. Its computer vision system will analyze footage from video cameras, identify the presence of moose, wild boars, foxes, and other animals, and send alerts to municipal services and drivers.
The alerts are expected to appear on electronic road signs and be broadcast over loudspeakers. Eventually, drivers may receive warnings before they reach a dangerous stretch of road.
An Animal on the Road Triggers an Alert
The neural network has been trained on thousands of examples. Developers fed it large collection of photographs of moose, wild boars, and foxes. The program gradually learns to recognize their distinctive features: body shape, legs, antlers, silhouettes, proportions, and patterns of movement. Once an animal is detected, the system automatically sends an alert through three channels: an electronic road sign, a loudspeaker, and a dispatch center. In the future, a warning could appear in a navigation app or arrive on a driver's phone before they reach the hazardous stretch of road. The software also has a confidence threshold. If the neural network is 90 percent certain, it triggers an alert; at 40 percent, the image is more likely to show a shadow or a branch.
The project is still at an early stage. The neural network must learn to distinguish individual animal species. The developers estimate that this phase will take at least six months.
The Cameras See More Than Faces – They See Foxes, Too
Russia already has experience using AI to recognize pets and monitor wildlife, and researchers and companies are also working on ways to detect wild animals in urban areas.
NtechLab, initially known for its facial recognition technology, later began applying similar algorithms to animal recognition. In the Moscow Region, the company launched a service called Poisk Domashnikh Zhivotnykh (Lost Pet Search). Users upload a photo of a missing cat or dog, and the AI-powered system compares it against databases to help locate the animal. By October 2025, the photo-search feature had been used more than 1,300 times.
NtechLab CEO Alexey Palamarchuk has said the company receives requests from Russia's northern regions to expand the technology so it can detect wild animals, such as brown bears, entering urban areas and immediately alert residents. A pilot project for a similar system was tested in the Khanty-Mansi Autonomous Okrug.
The Ministry of Digital Development of the Tula Region has also deployed NtechLab's AI-powered video analytics system. Its primary function is to detect violations such as people walking dogs in prohibited areas. During testing, the system spotted a fox in a city park.
In the Moscow Region, meanwhile, AI is being used to locate stray dogs. The number of reports and complaints about stray animals has fallen by 20 percent. Cameras are also capturing rare wildlife. Yandex Cloud and Yandex's School of Data Analysis developed a neural network that can identify ten different animal species in images and videos captured by camera traps, including snow leopards, also known as irbis. The system was tested in Saylyugemsky National Park in the Altai Republic.
The Kronotsky Nature Reserve and IT company CROC held the Wild Hack hackathon, which included a challenge to build a system that could automatically identify animals in camera-trap images. A student team developed a neural network-based tool that filters out empty frames and identifies animals in photographs, reducing the amount of time scientists have to spend reviewing images manually.
Avoiding Dangerous Encounters
Invian is already developing solutions for analyzing road traffic. In Kazan, its Incidents system automatically detects crashes, vehicle collisions with pedestrians, and other dangerous situations. The company plans to apply the same approach to detecting wild animals.
Since the beginning of this year, 191 traffic accidents involving wild animals have been recorded in Tatarstan. The technology could help reduce dangerous encounters between drivers and wildlife, particularly on highways outside urban areas, near forests, and on the outskirts of towns and cities.
The system is likely to be in demand in regions with large populations of big wild animals, including Siberia, the Urals, and the Russian Far East. Representatives from Argentina, Chile, and Brazil have already expressed interest in the development. These countries face similar problems with animals wandering onto roads.
