For the first time in Medaram Jatara in Mulugu district, the biggest human congregation witnessed every two years in the State, police are using analytic software to predict stampede-like situations and traffic jams.
The video footage of nearly 350 surveillance cameras installed in and around Medaram Jatara venue would be sent to a server over which the analytics are run in real time. The software will help police in management of people and vehicles at the massive Jatara.
Surveillance cameras were used earlier also. This time we are linking them with analytics software. A technical team from Hyderabad will come down to Medaram for this, a senior police officer associated with the security arrangements for the event told The Hindu. The technical team will be available at the special command control room at the venue and help police assess movement of devotees and vehicles by analysing the video footage.
Real time analysis
Taking density as parameter, the analytic software will help estimate number of people present in a specific area. It will also assess how many women, men, children and older people are there, the officials said. Though precise numbers cannot be given, the analytics would surely give 80% to 90% clear figures, they said confidently.
Running the software would also assist the police in assessing behaviour of the crowd. Normally, the figure of one person per square foot is normal. But if it crosses 1.5 person per square foot, it suggests an alarming situation and hints at possible stampede-like situation, the officials said. The authorities believe the software would be of utility for them while assessing the movement of people at the queue leading to the main venue.
A separate video wall is being set up at the command centre. Here video footage of the parking space, queue lines and main venue where people offer prayers is analysed. Interestingly, the police are planning to use facial recognition software to identify pick pockets and other offenders who try take advantage of the crowded situation.
Photos of persons indulging in pocket picking and stealing mobile phones would be integrated with the software. It will identify the offenders if they are seen in the crowds and pop up alerts.