Introduction to Video Analytics In CCTV Leave a comment

If you have an interest in CCTV or even just technology in general, you might have come across the term “video analytics”. The market for this is showing signs of increasing growth due to their uses in a wide variety of industries. But what exactly are video analytics and why is this technology so sought after? Keep reading to find out.

What Are Video Analytics?

Video analytics is defined as “a technology that processes a digital video signal using a special algorithm to perform a security-related function.” In other words, it is a technology that analyzes video footage to perform certain functions for improved security. This technology is extremely useful because it’s like having an extra eye to catch important events that you might miss from normal monitoring. There are 3 common types of video analytics: fixed algorithm analytics, artificial intelligence learning algorithms, and facial recognition systems.

Facial Recognition Systems

You might already be a little familiar with one of the video analytics, facial recognition, if you have a newer model of the iPhone. On the iPhone, the user can choose to set up face recognition to unlock their phone instead of having to type in a pin or use their fingerprints. Like fingerprints, every person has a unique facial signature. Even though you might resemble someone, no one’s face is completely the same. Face recognition identifies different people based on the mathematical measurements of their facial features.

In CCTV, face recognition works by matching up the faces captured from CCTV footage to a predetermined database of registered people and their photos. This video analytic in CCTV is useful for many purposes. For example, if a store owner registers a repeat shoplifter into their database, an alert can be sent to notify them whenever the shoplifter enters to prevent theft. Additionally, face recognition can be used to recognize authorized personnel to only allow them entry into a worksite and keep out intruders.

Fixed Algorithm Analytics

Fixed algorithm and artificial intelligence learning algorithms can be confused to be the same thing because they have a common goal. Both of these video analytics are used to detect suspicious activity in CCTV footage, but how they do this is where they differ. Fixed algorithm analytics use specific algorithms to identify specific behavior. In other words, each behavior is capable of being detected due to a corresponding algorithm. They’re called fixed algorithms because each one is designed to look for a specific action.

There are many different fixed algorithms that are made to detect different behaviors, but common uses of this video analytic includes:

• Line crossing, or movement crossing over an imaginary line which may indicate intrusion
• People counting
• Detecting direction of movement (useful for identifying cars driving on the wrong side of the road)
• Detecting objects being left idle in certain areas (can be used to identify people loitering around a business or suspicious luggage that has been abandoned in an airport)

Artificial Intelligence Learning Algorithms

Artificial intelligence (AI) learning algorithms is a video analytic that functions through learning. Instead of being programmed with specific algorithms for certain events, they come as a blank slate. They are then fed information and learn how to detect things through a process of trial and error. We’ve previously covered one example of this, object recognition. With object recognition, the machine learns how to identify and recognize objects by being continuously exposed to different things. Object recognition is useful to spot potential dangers like weapons or to help with tracking the movement of an object.

Another way that AI learning can help improve security is by learning what is normal at certain hours in certain places. If something is detected that is not consistent with the past recordings, an alert will be sent for further investigation. In one given example of this, an airport installed a camera to watch for people getting onto the luggage carousel. However, a man was detected to have been picking up small luggage items and putting them into a larger luggage. The AI learning algorithm noticed that this was an abnormal event so authorities were notified and it was found out that this man was stealing baggage this way. Fixed algorithms couldn’t have detected this event because it wasn’t programmed to recognize that, it was only possible with the system learning what is normal and what is not.

Wrap Up

Video analytics is a useful tool for CCTV because it provides the user with more security by detecting suspicious activity that might have gone unnoticed. However, this technology is not limited to security purposes anymore. Now, retailers and businesses are starting to use video analytics to help them gain more information and insight about their operations. Video analytics is a fast growing industry due to their versatile applications and it’ll be interesting to see what else they will be applied to.

As always, don’t hesitate to contact us if you have any more questions. You can call us at 877-926-2288 or connect with us on social media
Facebook
YouTube
Instagram
LinkedIn
Tumblr