The helmetless rider menace is widespread through most Indian cities. Some of the most common excuses are that it ruins a hairdo, or is ‘inconvenient’, parroted by people who disregard basic safety.
Well, if you are a helmetless rider in Hyderabad, your luck might run out quicker than you thought. The IIT and the cops are working together to implement an AI system that will catch helmetless riders and fine them accordingly.
According to the Economic Times, the IIT and Hyderabad City Police have signed a MoU that will give the institute access to video data from the city’s CCTV network. The ready-to-be-deployed technology, is patent-pending.
This is how it works:
Installed partially in cameras and servers of the central police control room, the software is embedded on a card attached to CCTV cameras; it helps detect violators by sending out an alert to a central database.
The solution uses convolutional neural network technology (applied to analyse visual images), which primarily uses AI to mimic the human brain. The system can be put to use to detect other traffic violations, like zigzag bike driving, tripling, etc.
One of the research scholars, Dinesh Singh, explained that the system would be fully-automatic, and have a web interface to verify the alerts by the operators, traffic police, etc. From there, the system would connect to the existing RTO website to generate challans, and send a notification to riders via SMS.
Associate Professor at IIT Hyderabad, C Krishna Mohan backs the innovation, with experiments on sparse traffic in their campus, as well as dense traffic from the Hyderabad CCTV network. The results have been favourable, and the complete software can be deployed for real-world use.
The solution can be deployed at the entrance of cities, at toll gates for instance, or at select checkpoints and road intersections.
The institute hopes to get an industry partner on-board for making the technology commercially available.
(Edited by Shruti Singhal)