Early warnings to protect against attack

As part of a larger security refurbishment program for Pierce College, part of the Los Angeles Community College District, the challenge was to develop an intelligent security surveillance system that could quickly identify and track potential intruders.

In this environment with more than 100,000 students, shorter response times and improved data available to security staff and police would help limit casualties.

“With conventional camera images, an armed intruder is difficult to identify in a crowd,” explains security specialist Louis Barani who joined AECOM from his role as Director of Security at the World Trade Center in New York City. “The new system learns what normal movement looks like, for example students moving between lessons. When it picks up unusual movement, such as a group of students suddenly running or falling, the system user and the police are immediately alerted and have a view of the attacker. A digital profile of the attacker is then created, which enables them to be picked up on other CCTV cameras and tracked around the campus. Dangerous incidents can be stopped much faster than was previously possible.”

The solution integrates two software applications Giant Grey (formerly BRS Labs) AI-Sight and Qognify (formerly NICE Systems) Suspect Search. The Giant Grey behavioral pattern recognition software memorizes patterns in its field of view using cognitive machine learning. When it observes an abnormal pattern, it alerts the operator who acts accordingly. In an emergency, the operator uploads the intruder images into Qognify. This software develops a digital profile of the intruder and searches the camera system. It can do this for multiple individuals simultaneously. This information is relayed in real time to the police.

Wider potential
Barani worked on earlier versions of this solution for the World Trade Center and a navy base in Qatar. Following the successful school installations, other clients have included a major East Coast transit system during the Super Bowl season. The system has many potential uses, for example, when a suspect is identified on a citywide CCTV system, their movements can be tracked using their profile throughout a wide area —  a task that would have taken many hours by conventional means. Further potential applications for this technology include airports, shopping centers, music festivals and tourist destinations — essentially, anywhere where there are crowds and the risk of someone wanting to cause disruption or harm. The software can also be used to help detect incidents such as fires or floods.