So much video, how do you search through it?

There is no dispute that security cameras for your business or construction sites will deter a crime. However, most people don’t realize that hours of video footage gets created, and tt’s a daunting to look through.

Imagine having to find a vehicle break-in amongst hundreds of cars in a parking lot. It’s easy to miss an event or simply give up because you have more important tasks to take care of. With a dedicated security guard who has the time to look through video, this task can be delegated. That being said, this task usually falls onto the lap of the facility manager, operations manager, or IT department.

What is motion detection for Security Cameras?

Every security camera system records on motion detection. Within the software, a region or area of interest is selected, and any pixels that change within that area is considered motion. Further, the areas that are not in the region of interest are ignored.

Milestone Motion Detection Setup

 

One of the problems with motion detection is the sensitivity. Setting it too low results in events being ignored, causing video to be missed. For example, a person walking in the distance may not change enough pixels to warrant the cameras to record.

On the other hand, if sensitivity is set too high, the cameras end up recording useless video.

It’s not fun to watch video of debris and trees blowing in the wind.

What is Video Analytics in Security Cameras?

Video analytics in security cameras is a big leap forward from motion detection. Motion detection looks for pixel change, whereas video analytics uses intelligence to identify people, vehicles, and other objects. This is done to make video surveillance more efficient, cost-effective and less prone to human error.

Video analytics can be used in real-time to identify a security breach or can be used to analyze pre-recorded video footage to identify patterns, trends or specific incidents.

The video can be sorted by time and date or by specific events such as people or vehicle detection.

Make searching through video easier

Searching through recorded footage using video analytics in security cameras is a breeze compared to motion detection. On a rainy night, motion detection will pick up hundreds of false alarms, whereas video analytics will reduce false detection by upto 90%.

A video search takes minutes when using video analytics compared to likely giving up if searching with motion detection.

One of the simplest applications of video analytics is detection on a fixed background. For example, someone climbing over a fence and into a construction site or yard. It will be inefficient for a security officer to watch the cameras all night to monitor for someone crossing the perimeter.

However, with video analytics, the camera detects a person or vehicle in a specified area, and can trigger an alert. This results in getting a security officer’s attention, who can intervene before the intruder causes any damage.

Get alerted to a problem, before it’s a problem

Another common application of video analytics is triggering alerts in real-time, as an incident occurs. Let’s say your facility has a large parking lot where you are having vandalism or garbage dumping problems. You know the problem occurs after hours when there is no one around. You can set up an alert to be triggered if a car stops there for more than 5 seconds between 9PM-7AM.

Are video analytics in security cameras accurate? Or are there a lot of false alarms?

Video Analytics in security cameras have come a long way over the last decade. Until the late 2010’s, there was way more hype about it than actual installs. The accuracy of video analytic cameras (i.e., false positives and missed alarms) has improved dramatically over the past few years.

Some manufacturers are better at marketing video analytic features than they are at providing positive results. That being said, if you stick with a good manufacturer such as Avigilon, one of the pioneers with constant positive reviews, then the accuracy level will be excellent.

A common barrier against using video analytics is customers often complain against the cost. Many can argue that video analytic cameras are most useful for outdoor locations where busy scenes will cause motion-based recording to produce too much useless video.

Do you really need to have a security camera with people detection for an office? There is likely going to be little to no false motion recording and if something happens it won’t take that much more time to search for video in an office setting. On the contrary, if you have an outdoor camera it can easily be triggered to record for animals, tree’s blowing in the wind, debris, cars driving on the road in the background, etc.

5 Most Popular Video Analytics in Security Cameras that every business can use

  1. People Detection: This is the simplest and most used video analytics in security cameras. A region or area is determined during camera setup and when a person is detected the camera records. The event is tagged as “Person Detected” and you can run a search for this event over a certain time frame, for example “Person detected over the past 24 hours”.
  2. Vehicle Detection: Similar to people detection, this is also a commonly used. The event will be tagged as “Vehicle Detected” and searches can also be categorized by vehicles.
  3. Loitering Detection: The difference with loitering detection is that a time frame can be set. The camera can ignore if people or vehicles are detected in an area until they are present for a certain time frame, for example 5 seconds.
  4. Direction Violation: This video analytic is useful to detect if people or vehicles are moving in the wrong direction. A warehouse that requires people or vehicles to move in a certain direction can use this feature to ensure safety procedures are being followed.
  5. Camera Tampering: This is a great video analytic that can be used to detect if the camera is being tampered with. When the scene unexpectedly changes, for example it is knocked down, an event can be triggered.

The Next Big Thing for Video Analytics – Deep Learning and Artificial Intelligence

The next big thing for video analytics in security cameras is Deep Learning and Artificial Intelligence.

Many manufacturers are releasing security cameras and video management software that can perform facial detection and unusual motion detection. Although this technology is still in early stages, it will just continue to improve.

Security cameras with facial detection software can now open doors, for example, when an authorized person is detected. Features such as Avigilon’s Appearance Search will allow you to search for a male wearing black pants or a woman wearing a red dress.

Unusual Motion Detection is another example of artificial intelligence built-in to Avigilon security cameras. It will continuously learn the regular activity in your facility and flag events that seem unusual.

Video analytics in security cameras will make your video surveillance system easier to use. According to the world’s leading video surveillance information source, IPVM, 2020 is the year video analytics will go mainstream. Is your business ready to upgrade?

gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==
Written by : Carlo Di Leo

At the age of 24, with no experience in the security industry or any money in the bank, Carlo quit his job and started Spotter Security from his parent's basement. Founded in 2004, Spotter grew from a single man operation into a multi-million dollar security system integrator that caters to businessess and construction sites across Canada.

Contact Us

Free Up Your Time To Get Back To Your Most Important Work