As do many other cities, Jyväskylä is actively developing the measurement of various modes of transport. Analyses are now done also with the help of a solution based on artificial intelligence.
Innovative solutions for the measurement of traffic are being developed all the time. One example of such development is the monitoring of mobile terminal devices, which is a handy tool for the monitoring and interpretation of, say, bigger mass movements. Google, for one, offers data on the smoothness of traffic and on travel times between various points. All this based on – you guessed it – the movement of mobile terminal devices.
Even so, when the aim is to count traffic volumes in precise figures and identify various modes of transport, you need counters installed on the road or street in question or in its vicinity.
The latest advancements in this respect include the exploitation of neural networks and deep learning, or what is also referred to as artificial intelligence. In this solution, the AI software is taught to identify all traffic shown in a video feed. The AI identifies and counts the vehicles, categorising them as well, at the same time. The algorithm simultaneously identifies and categorises all pedestrians and cyclists as well as the direction and speed of the subjects. All this from a video feed produced by a single camera.
An AI identifies all modes of transport from a single display of video feed.
JYVÄSKYLÄ TAKES A LEAP TOWARD AI MEASUREMENT
Jyväskylä took its first step toward the new era of AI in traffic measurement.
In cooperation with Nodeon, the city selected one of the city’s busiest streets as a challenging measuring target. The street links the extensive residential areas east of the city centre to the modern employment and residential area of Lutakko, the waterfront street that passes the city, and the city centre with the help of the impressive Kuokkala bridge. The goal, then, was none other than counting all traffic on the bridge.
This would have been a challenging task with traditional measurement techniques. The traffic on the bridge is two-way, in addition to which there are two busy lanes for pedestrians and cyclists on both sides of the street. With traditional technologies, the solution to the problem would inevitably translate into a combination of various techniques and induction loops that would need to be sawed beneath the pavement – which is not usually very easy on a bridge.
Nodeon approached the challenge with the help of a single camera. The device analysing the video feed – placed in the measuring point and equipped with AI – analyses the traffic on the bridge on a continuous basis, identifies and categorises the vehicles as well as the pedestrian and cycling traffic, and transmits all of the data in real-time to Nodeon’s centralised Smart City storage and analysis environment for measurement data.
AI offers a genuinely credible, in some situations even superior, alternative to traditional measurement techniques!