“If I had known that I would have been caught, I wouldn’t have gone.”
Those were the words of a 31-year old suspect known simply as Ao, who was arrested by Chinese police officers after being picked out of a crowd that was 60,000-strong during a music concert on April 7th.
How did they do it, you ask? Well, this was a prime example of China’s facial recognition technology working at its best. In this particular instance, Ao’s face was picked up by CCTV cameras positioned at the entrance to the concert and was immediately and automatically flagged by the system as a suspect who was wanted for unspecified “economic crimes.” Police officers responding to this were able to pinpoint his exact location within the stadium, right up to the row where he and his wife were sat enjoying the festivities.
Here’s another scenario which played out in Rotterdam, outside a jewellery store to be precise. A man walks up to the store entrance and attempts to push the door open, but nothing gives. As it turns out, the man in question happened to be a known shoplifter with a criminal record, and his face had been picked up and flagged by facial recognition cameras positioned at the entrance. In this case, the cameras were integrated with the store’s access control systems to automatically lock out known offenders, which worked perfectly here.
More recently, facial recognition was used to rescue a three-year old child who had been kidnapped in Shenzhen, China. This technology, and video analytics by extension, are widely regarded as the future of surveillance, crime detection and prevention. In fact, given the incident in Rotterdam was a pilot project carried out seven years ago, we might well argue that the future is already here.
This is why we recently partnered with Axxonsoft, a Russian software developer whose solutions currently account for over 70% of the Russian market, to introduce video analytics to the Kenyan market. This partnership is part of our continued commitment towards innovation and will add to our long list of firsts in the region.
In truth, facial recognition is only a small element in the wider video analytics umbrella. Our systems are capable of identifying subjects regardless of their facial hair, hairstyle, eyewear or aging. Beyond this, the system will also flag known suspicious characters whose images have been fed into its database and send a notification to you in real-time.
Now, CCTV surveillance footage has often been used as evidence to put many investigations to bed in the past. What most don’t often see is detectives combing through hours or days of footage to pick out a suspect’s face for instance, which gets infinitely more difficult in high traffic buildings or public places. However, with our solution, one can simply feed an image of the subject’s face into the system and let it do all the work for you. It will pick out snaps of the subject every time (s)he was captured on camera within a pre-set time frame, making such searches take a matter of minutes rather than hours or days.
As far as crime detection is concerned, the system can also be configured to send an alert in the event of:
- Motion within a particular zone.
- Stopping within a particular zone.
- Too many objects or people within a particular zone.
- A subject leaving a particular zone.
- Suspicious abandoned objects (like a backpack)
- Loss of signal.
- Positional change.
- Noises (like gunfire or a car alarm)
In some ways, it could also be argued that this surveillance solution predicts crime. Take the “Tag and track” feature for instance. The system can be configured to pick out certain parametres, such as a man in an overcoat, raise an alert then simply track the subject’s movements within the premises.
In this way, one can track a suspicious character’s movement within a building, a supermarket for instance, meaning you always stay one step ahead.
In the case of Ao for instance, when a known offender steps into your premises, the system can give you up-to-the-minute information on their exact location, making it easier to nab them using this feature.
And it’s not just faces that the system can pick out. It’s also capable of Licence Plate Recognition. Here, the system picks up the licence plate details of one vehicle every three seconds, catering for vehicle speeds of up to 250 kilometres per hour. Alternatively, one can search the footage archives for a particular number plate and find this with ease.
As far as analytics go, features like the Visitor Counter Detector and Queue Management only scratch the surface of our solution’s capabilities. You can, for example, tell how many people came by a particular checkpoint within a certain timeframe. You could even tell, in the instance of a supermarket, which aisles are the most frequented or which attractions at an exhibit had the most attention.
In a day and age where the online sphere is awash with cyber threats, the security of your information is key. This is why the “Failover” system is so important. Your data is hosted on secure servers with the option to configure a reserve server during configuration. This means that should the main server be compromised in any way, the standby then kicks in.
Simply put, video analytics is a technology whose time has come!
Interested in how this can help you secure your business? Talk to us today!