PRAYAGRAJ, Sept. 20 -- : After the stampede that claimed several lives during the Mahakumbh 2025 at Sangam earlier this year, as many as 50 high resolution, AI based infrared cameras to monitor pilgrim influx during the Magh Mela at Sangam have been proposed to be installed under the Smart City initiative. The measure is expected to help in preparing a roadmap for a mishap free Kumbh 2031. With an aim to assess the arrival of pilgrims and tourists at Sangam post Mahakumbh, a joint survey was carried out recently by Prayagraj Mela Authority alongwith Prayagraj Smart City Limited (PSCL), Project Management Consultant (PMC) and Master Systems Integrator (MSI) modules at key locations for assessing human count and vehicle count. During the survey, post-Kumbh effect was clearly visible as despite ongoing floods, in August, a surge in visitor count was witnessed on key religious days like last Monday of Sawan (August 4) which recorded high pilgrim movement at Kali Ramp (3,335) and Triveni Ramp (5,551). This surge indicated a robust post-Kumbh pilgrim influx. Further, unlike earlier, on non-religious days like Independence Day (August 15), pilgrim movement was recorded at Kali Ramp (790) and Triveni Ramp (3658) which also reflected sustained pilgrim interest even during a national holiday, a notable rise compared to typical post-Kumbh periods in previous years. According to IT Manager, Smart City Yojana., Prayagraj Mani Shanker Tripathi, presently four AI based, high resolution infrared cameras are installed near Triveni ramp and Kali ramp at Sangam while more AI based cameras would be made operational post monsoon at three other locations, Sangam nose, New Kali Road and VIP Ghat. "These high resolution cameras are AI enabled, capable of accurate detection of the count of pilgrims and vehicles arriving in the mela area, both during the day as well as night time. The post Mahakumbh effect is clearly visible. Unlike previous years, this year after the mela, a significant count of pilgrims has been recorded every day," he added. As per the expert, the Al Model Architecture is aimed at providing real-time detection of pilgrims and vehicles, optimized for high accuracy in crowded environments. The algorithm tracks objects across frames, preventing double-counting and ensures precise counts while during data preprocessing, frames are pre-processed to handle lighting variations and noise, enhancing detection reliability. Official records revealed that the performance metrics of the exercise achieved a mean Average Precision (mAP) of 0.92 for detection and 95% tracking accuracy. It may be mentioned that under the initiative, installation of a total of 50 AI based cameras at key locations in the mela area was on cards for the forthcoming Magh Mela, Kumbh and other occasions. Presently 82 CCTV cameras are installed at 13 different locations in the mela area, feeds of which are monitored through Integrated Command and Control Centre (ICCC)....