Cameras are the eyes of any security system. They can record everything that is happening 24/7/365. They are constantly collecting and recording data. However, basic cameras are pretty passive. They’re good for reviewing an incident after it happens and they can deter crime by their mere presence, yet cameras alone cannot stop or prevent crimes from occurring. What if you could give the eyes of your security system a brain?
What if they could learn from their surroundings and alert you when they sense something suspicious or out of the ordinary? This capability is no longer hypothetical or science fiction. It’s real and it's being used now with its use continuing to grow dramatically.
New Technology Allows Cameras to Learn Faster
Video analytics and smart cameras have been around for a while. Cameras can recognize when someone is moving in the wrong direction, entering a restricted area or running through a corridor. In retail applications cameras can even tell if someone is spending an unusual amount of time in a store aisle or if a shelf of products suddenly disappears. These smart cameras can send alerts to security and operations professionals warning them that something suspicious has happened. Cameras definitely got smarter, but they were still limited in how much they could learn.
Recent dramatic advances in technology now allow software in cameras to analyze large amounts of video very quickly. This ability has put machine learning on steroids and allows cameras to learn much faster. One example is facial recognition. Cameras can now pick faces out of a crowd and match them to other images in data bases. This is extremely useful because it allows cameras to identify a suspicious person in a crowd or building in real time.
Another example is license plate recognition. LPR allows security personnel or law enforcement to take an image of a vehicle license plate and immediately get information on ownership, past citations, year, make and model. This is valuable information that can be used in the case of an emergency or during an investigation.
As cameras learn faster and get smarter many security professionals are looking at a future where predictive analytics are commonplace. If cameras can learn the behaviors and movements that happened before an event occurs they could help prevent crimes. A Japanese camera maker claims it has cameras that can predict when a person will shoplift. By analyzing video of shoplifters the manufacturer says the camera can now detect the body movements that happen prior to a shoplifting incident. When the camera senses someone is going to shoplift it sends an alert.
The company is still looking to perfect the technology and they have had false alarm issues with people restocking shelves and indecisive shoppers. But, it looks like predictive analytics are well on the way to becoming a reality.
As AI becomes more common and security systems get smarter the security industry will have to deal with the privacy issues that follow. There will have to be regulations on the collection and storage of images. It will take working with public and private organizations to come up with laws that will help keep people safe while protecting privacy.