Three IT Graduates Develop AI-Powered 'Traffic Eye' to Assist Traffic Police With Real-Time Road Surveillance

 

Kathmandu. Three recent information technology graduates have developed an artificial intelligence (AI)-powered traffic monitoring system aimed to help Nepal Police improve traffic management and road safety.

Rohan Khanal, Shail Malla, and Shubham Kadaria created the system, named Traffic Eye, after participating in the Hackathon for Safety organized by Lumbini Province Police from June 13 to 15.

The three-member team built the project within the hackathon's 36-hour development period under the public safety and traffic management category.

According to team member Rohan Khanal, the system is designed to assist traffic police by automating road surveillance rather than replacing personnel.

"Traffic police currently rely on CCTV cameras, but monitoring the footage often requires manual observation. Our AI system can continuously monitor roads 24 hours a day, seven days a week," Khanal said.

The prototype currently includes three key features: detecting helmet violations, identifying and capturing vehicle number plates, and sending real-time alerts when traffic congestion increases.

Following the hackathon, the team has continued developing the project by adding new capabilities. One of the major upgrades underway is the integration of Optical Character Recognition (OCR) technology, which will automatically read vehicle registration numbers from captured images and convert them into searchable digital text.

According to Khanal, this would allow authorities to quickly search databases for specific vehicle numbers, making it easier to identify traffic rule violators and trace lost or stolen vehicles using CCTV footage.

The developers are also training the AI model with larger datasets to improve its accuracy while enhancing the software's backend and dashboard to help police analyze traffic data more efficiently.

Beyond traffic management, the team believes the system could also detect road accidents, physical altercations, and the movement of wildlife along highways near national parks, enabling faster responses from authorities and improving the safety of both people and animals.

Despite the prototype's progress, Khanal said large-scale deployment would require significant digital infrastructure, including GPU-powered servers, high-capacity data storage, and integrated software systems capable of processing CCTV footage in real time.

He added that institutional support, particularly from the Nepal Police, would be essential to deploy the technology for everyday use.

"If implemented, AI could continuously monitor CCTV feeds, allowing police personnel currently assigned to watch surveillance footage to focus on other traffic management and enforcement duties," Khanal said.

पछिल्लो अध्यावधिक: असार २२, २०८३ १४:२५

टिप्पणीहरू