Researchers at Ohio State University use image analysis with the help of Deep Learning
In addition to cameras in buses that monitor the interior, the traffic situation can also be observed and evaluated. If traffic is too heavy, variable message signs can be activated to reroute vehicles to avoid congestion. The system was developed by researchers at Ohio State University (https://www.osu.edu). It uses deep learning software to identify parked and moving cars as well as other road users.
Extremely precise system
According to the scientists, the algorithm can also project real coordinates of the road network from a bird’s eye view by using satellite navigation data and information from digital maps. The system is said to be so precise that it can also detect a bus straying from its planned route, says electrical and computer engineer Keath Redmill. The cost of the cameras is minimal, he says.
With the cameras already installed on Ohio State University campus buses, the researchers are demonstrating that they can automatically and accurately detect the number of vehicles on urban streets, recognise objects on the road and distinguish parked vehicles from moving ones. In previous research, the researchers had already found that using mobile cameras provides much better spatial and temporal coverage than using sparsely and often temporarily placed sensors that do not have visibility of many streets and roads in a city.
More efficient transport system
“If we collect and process more comprehensive, high-resolution spatial info about what is happening on the roads, the efficiency of the transport system can be effectively improved,” says Redmill. And colleague Rabi Mishalani adds, “If we can measure traffic in a way that is as good or better than what is traditionally done with fixed sensors, then we have created something incredibly useful extremely cheaply. Our goal is to build a system that can do this without a lot of manual intervention.”
If the system is widely adopted and integrated, the expansion of traffic routes could be better planned and existing ones used more effectively, they say. Such advances in traffic control could bring shorter travel times and greater travel choices. “Transport planners, engineers and operators are making important decisions about the future of our roads, so as we design transport systems that will work for the next 30 to 50 years, it is vital that we provide them with data that will enable them to improve the efficiency of the system and the level of service for travellers,” Mishalani says.