Video content analysis | Occupational safety
Video content analysis (VCA) is changing the face of CCTV security, says Andrew Hunter, director of products and solutions at Chubb Fire & Security, in the photo.
There is no doubt that video surveillance has revolutionized the world of security. Since its first use in the 1940s, it has evolved enormously. Today, with the emergence of video content analysis, video surveillance as a security tool has become powerfully intelligent and is changing the face of video surveillance.
Video Content Analysis (VCA) has taken CCTV security to a whole new level. It gave security professionals the ability to assess events, identify vehicles, people, animals, and monitor significant moving events, while canceling unrelated events such as moving trees and branches.
While “motion detection” isn’t something new – it’s been around for 30 years or more – video content analysis is much newer. Over the past five to ten years, this technology has been incorporated into CCTV cameras or Network Video Recorders (NVRs) and allows the system to work automatically to detect and analyze temporal and spatial events. More recently, additional analytics can be applied after cloud recording to allow the application of artificial intelligence to constantly evolve, refine and improve a security system.
Originally, video content analysis was mainly used in high-risk or high-security environments, but due to cost reduction, it has become much more widespread and is used in all applications, scenarios large-scale commercial and industrial to people. houses.
Video data generated by IP CCTV cameras gives security teams better situational awareness. They can use the data to process, categorize and analyze objects and activities. It can identify temporal shapes, spatial events, and even direction of travel. These powerful features dramatically change the face of video surveillance. At Chubb, the level of security we can deliver is far more powerful, accurate and efficient than ever before, and the pursuit of incremental improvements is unapologetic.
In a real-world scenario, for example, video content analysis can identify a security breach by analyzing a person’s movements. It can mean the difference between a routine swipe of an ID card at the front desk and an intruder jumping a barrier or breaking down a door.
The system is also smart enough to determine security vulnerabilities through facial recognition. For example, an alarm would go off if an authorized person walked through a door but was immediately followed by an unidentified person.
Video content analysis allows us to solve our customers’ most important problems much more effectively. If an incident occurs, we no longer need to spend a lot of manpower or time searching through the footage of the event. The system will find the actual event almost instantaneously because it “marks” the moments by type of analytical event. With this automated system, we also eliminate any human error that could potentially occur, such as missed details that can impact incident response time.
The automated system also gives us a much higher degree of accuracy in determining whether an incident is real or not. Our central video monitoring station was able to reduce the number of false alarms by up to 80% by implementing video analytics. Assessing confirmation of the validity of incoming alarms and presenting only genuine people or vehicle activations to the operator is much more efficient for the customer and for us.
The other benefit of using video content analysis is a reduction in the amount of hard drive storage you need. The system will only record and retain an authentic event, which ensures that the storage is not filled with unnecessary hours of video content. The reduced need for hard drive space can mean fewer servers, resulting in lower power consumption, saving energy and money. We find that power consumption and storage space requirements become a key priority for many of our customers when selecting their system.
While CCTV is synonymous with security, video content analysis can be used much more widely. For example, many of our customers use mobile phone notifications to notify them of relevant events that require immediate attention. This is particularly useful as it allows the customer to continue the daily workload while receiving a global, real-time view of all of their activity and all the interventions they need to attend.
For example, one of our customers, a large cattle producer, uses testing to monitor his cattle as they go into labor. As work can take place at any time, day or night, the farmer can set up an alarm when movement is detected in a cow. Cows start circling before calving, so when this starts to happen the farmer can be with his cattle to tend to them at the critical time and not have to stay up all night in anticipation .
More and more solutions rely on multiple analytics providing a series of events that would trigger an alarm. Last year, Chubb provided a major petroleum retailer with a platform that recognizes if people are on the forecourt and are dispensing fuel safely. This was achieved by multiple analyzes applied simultaneously, from smoke detection to motion and pump impact alerts.
As I mentioned at the beginning, CCTV technology is constantly evolving. Over the next two years, I expect to see further advances in self-learning video analysis. This will help increase the accuracy of false alarm detections and present only true alarms, allowing organizations to respond more effectively to different scenarios.
Rapidly evolving capabilities will also be an important feature in the years to come. During the pandemic, video content analytics allowed companies to count the number of people who entered their buildings and detect if they were wearing a mask, all designed to minimize the spread of covid-19.
Should other unknown requirements become necessary in the future, I’m sure other iterations of these scans will become available during development.