Samara Region Turns to AI to Track Pit Bikes
AI-powered cameras keep a close watch on road conditions, detecting a wide range of traffic violations. They can identify drivers who aren’t wearing seat belts, motorists talking on their phones behind the wheel, and motorcyclists riding without helmets. Now, neural networks can also “see” pit bikes on public roads. Samara Region is the first in Russia to roll out this technology.
The solution was developed by Russian company Netvision. Connected to the Safe City system, the service analyzes video feeds from urban surveillance cameras. In the first three weeks of its pilot operation, the neural network helped identify more than 500 pit bike trips.
A Network of Smart Cameras
Why are pit bikes drawing attention? These compact motorcycles are designed primarily for motorsports on dedicated tracks and in closed areas. Many models lack the equipment required for use on public roads, and riding them there poses risks to both riders and others. According to Russia’s Interior Ministry, 288 traffic crashes involving children riding pit bikes occurred in the first six months of this year. Ten children died and 235 suffered serious injuries. Over the past two years, the number of incidents involving underage drivers of motorized vehicles has risen by 68%.
The neural network will distinguish these compact motorcycles from cars, conventional motorcycles and other road users, even in heavy traffic. The system receives images from street cameras, identifies objects, records each detection and links it to a specific location in the city’s infrastructure. The data is then accumulated on the platform and forwarded to the responsible agencies. This will allow the neural network to build statistics showing where pit bike riders are most frequently spotted.
The Road to a Digital Future
In recent years, Russia has launched a range of video analytics projects in which neural networks analyze camera footage and automatically detect events on roads and in urban areas. MTS has deployed a similar system in Sirius. There, AI-powered cameras detect violations of electric scooter rules, including riding across pedestrian crossings without dismounting, traveling in prohibited areas, carrying two people on one scooter, and improper parking. The information is sent to the scooter-sharing operator, which then notifies the user. According to the company, more than 90 percent of users who received a warning tried not to repeat the violation.
Moscow’s Traffic Management Center also uses neural networks to detect violations by electric scooter riders. During the 2024 season, rental operators issued users 132,000 fines and blocked 127,000 accounts. Meanwhile, NtechLab introduced a system designed to detect traffic violations by cyclists and bicycle couriers. It is intended to recognize specific types of violations, such as riding on a sidewalk when a bike lane is available, and issue voice warnings.
Many intelligent video surveillance systems do not operate independently. Instead, they are integrated into regional Safe City hardware and software platforms, which combine camera networks and data processing centers with specialized analytics modules. Samara Region is no exception. Its platform runs detection systems for pit bikes, electric scooters and bicycles side by side. Over the past month, the system recorded more than 11,000 violations by electric scooter riders and more than 5,000 by cyclists.
A Tool, Not an Inspector
The system does not replace a traffic inspector. It does not automatically issue fines or identify individual riders either. At this stage, its role is to detect vehicles, build up statistics and pass information to the relevant agencies, which assess the situation and respond when necessary.
The developer says the data can help officials assess the scale of the problem and make a stronger case for future traffic management decisions. The company plans to further train the neural network to recognize enduro and motocross bikes and expand the technology to other regions. Netvision’s platform is already listed in Russia’s Unified Register of Domestic Software and is used at government and commercial facilities across the country.
