How to Leverage the Data from Security Surveillance Footage as Video Evidence
The Expanding Video Surveillance Landscape
Nowadays, video surveillance cameras are ubiquitous and commonplace, appearing in environments from retail stores to public schools, college campuses, hospitals, parking garages, banks, corporate offices and transit stations. It’s a great advantage to extend human sight and keep an “eye” on multiple areas – especially crime hotspots. Some facilities maintain a security staff who monitor video feeds in real-time, but in many cases there is far more video footage than security staff to watch it. The transit authority of a large city, for instance could have tens of thousands of video cameras in its network. Monitoring the recorded video in real-time would require significant manpower and – even if there were enough security staff to view a network of camera feeds — the threat of human error looms large: It would be easy for a human monitor to overlook or miss important details within the footage.
Manual Video Review is Prone to Error
Since real-time monitoring is often impractical and unsustainable, most video footage is reviewed only on an as-needed basis to drive post-event investigation for law enforcement or security staff. Even in the aftermath of an incident, the task of sifting through minutes, hours or days of footage to find clues is time-consuming and prone to human error or distraction. Video evidence is a powerful tool for law enforcement agencies worldwide; however, more often than not manually reviewing the footage proves too time-consuming and work-intensive. All the while the likelihood of human error or distraction looms. Unfortunately, however, that leaves much valuable video surveillance footage underutilized.
Video Content Analytics Increases Accuracy and Efficiency
This is why more businesses, institutions and municipalities are deploying video content analytics technology that processes video, identifies objects (such as people or vehicles) in the video footage, and classifies and indexes the video object metadata so that footage can be easily and quickly searched and analyzed. This allows dispatchers and officers to scan hours of footage in only minutes, and pinpoint objects and persons of interest by searching according to specific criteria. This filtered search functionality enables rapid and accurate video review so that police can gather key evidence and make better and faster decisions. After a security incident – or as a situation develops in real-time – time is of the essence, and the ability to accelerate video search and response to unfolding events is critical for helping law enforcement and security personnel track down suspects and provide key evidence.
Solving Crimes on Campus with Video Analysis
In 2018 the University of Pennsylvania began using video content analytics to solve crimes on campus. The student newspaper, the Daily Pennsylvanian, describes how video intelligence technology enables users to filter camera footage for objects of interest such as men, women, children, and vehicles, with attributes such as appearance similarity, color, size, speed, path, direction, and dwell time. The technology “greatly reduces the time detectives take to comb through the footage.” The university credits the technology with driving at least four arrests: “two burglaries, one theft, and one bicycle theft. One of these burglaries was committed by a serial burglar of office spaces. He was arrested for 59 burglary and trespassing-related offenses shortly after his spree began.”
With video intelligence software, security and law enforcement agencies have multiple ways to collect intelligence and evidence from video data. It is a logical next-step technology for organizations that already invested in video surveillance systems, enabling them to maximize existing investments in video.