A street camera. (Representative image)
Credit: iStock Photo
Bengaluru: Bengaluru Police are adopting technologies to stay ahead of the new-age criminals in cyber and other rapidly evolving domains. While the thrust is on pre-empting new and emerging forms of transgression, technologies are also steering efforts to check a more immediate, real-time threat – street crime.
GPS and auto-dispatch systems have helped the Hoysalas improve their response time. Now, the Police Department is looking at technologies that can help the force predict, and prevent, crime. Under the Safe City Project, 7,500 high-resolution cameras with facial recognition features have been installed, with a command centre accessing visuals from these cameras. At present, the police have a licence to incorporate facial recognition features in over 1,000 cameras.
A senior police official notes that socio-economic factors continue to drive street crime but as the focus shifts to counter-cybercrime strategies, it is also important to engage technologies to address the more conventional forms of crime.
An AI-run Facial Recognition System (FRS) is set up at the command centre and linked to a Karnataka State Police database that has case details and facial features of people with criminal backgrounds. When a person recorded in the database enters the range of any FRS-linked camera, the system flags this presence to the command centre. Once the face is recognised and marked, the system displays the person’s criminal records; subsequently, an officer monitors the live movement of the subject.
Cameras installed under the Automatic Number Plate Recognition (ANPR) system read vehicle registration plates and provide similar access to the command centre.
Officers at the centre can pin any individual or vehicle on the system to track live movement within the cameras’ range, providing the last known location of the subjects being monitored.
In a pilot project launched in June, 12 cameras installed at the Majestic Bus Stand were linked to FRS. On average, these cameras detected 200 persons with criminal records, every day. A Rapid Action Team from the Upparpet Police Station responded to these detections and apprehended about 10 wanted suspects.
Several police officials DH spoke with pitched FRS and ANPR as highly effective tools in countering real-time crime. “Unlike traditional methods that require checking all the cameras, FRS involves a streamlined, real-time process. The system immediately provides the subject’s last known position and enables closer tracking,” says Satish, an investigator at the command centre.
Predict and prevent
Police are also adopting predictive action as an effective strategy, by utilising call data from the 112 helpline, compiled over five years. These recordings help the officers map the origin, time, type, and gravity of the crimes.
Every day, this data is processed using analytical AI tools to predict the likelihood of crime at specific locations. Based on the findings, investigators at the command centre direct the Hoysalas to patrol areas with a higher probability of crime.
Investigators contend that expanding FRS can bolster efforts to control street crimes – particularly, the ones reported as soon as they occur – and help restrict crimes in crowded areas through continuous monitoring.