Video surveillance has been around for a very long time — probably much longer than most people realize. Closed circuit television (CCTV) monitoring has been used since the mid-1960s. A major hallmark came in 1969 when New York City installed cameras on its Municipal Building. As CCTV was more widely embraced, it probably didn’t take long for users to realize that monitoring the CCTV feeds was no easy task.
Research shows that as the number of cameras an operator is tasked with monitoring increases, their ability to do so effectively decreases. So while CCTV could provide comprehensive views of an area’s activity, this information wasn’t being leveraged to the fullest — not even close.
Eventually a solution to this problem emerged, in the form of video content analysis, now more commonly referred to as video analytics (VA). Although the technology was promising it has taken several years to be refined, fine-tuned and tweaked such that it is finally ready for prime-time deployments. Along that sometimes bumpy and contentious route it has had to also overcome false starts and being overhyped, as well as being cost prohibitive. Find out how VA advances are alleviating false alarm issues while enabling a wider variety of applications.
False Alarm Failures Fading Away
Using a combination of algorithms, video analytics analyzes captured video in real-time and presents alerts about whatever the application is programmed to identify. In early versions this was primarily motion. But as you can imagine, things move all of the time, especially outdoors. As a result, the first uses of VA applications resulted in a high number of false alarms.
While these systems did manage to identify suspicious movement, they would also notify operators of unintended situations, such as when the wind picked up or when a vehicle’s headlights blurred the scene or when it started to rain. Unfortunately, too many false alarms eventually caused security operators to ignore alerts, which defeated the purpose of having video analytics in the first place.
Fortunately, VA technology has come a long way. During the past 10 years, it has evolved at a rate similar to other technologies, and is now in its fourth generation. Today’s VA applications are able to do much more than just identify motion, and false alarms have been reduced to negligible rates.
For example, VA applications are now able to automatically filter out motion caused by wind, snow, rain and change of lighting. Some applications now also have the ability to detect tampering, and can automatically adjust the visual parameters of enabled video cameras according to individual scene characteristics to ensure optimal brightness and contrast for video viewing and recording.
3 Breakthroughs Open Up Uses
The more recent evolutions in VA applications have made them exponentially more effective. Three particular technological advancements have really bolstered reliability and opened up possibilities: Object classification; pan/tilt/zoom automatic tracking; and multisite dashboards.
Object classification — Most basic VA applications simply detect moving objects but don’t distinguish their nature. Using object classification, newer video analytics software is able to differentiate among different types of moving things, such as a human or a car. It can also filter out certain mobile elements, such as moving vegetation, a shaking fence, shadows and car lights, so they will not set off an alarm.
With the ability to classify objects, VA applications are able to better identify true potential threats, and of course rule out false alarms.
Pan/tilt/zoom automatic tracking — Pan, tilt and zoom (p/t/z) cameras have the ability to follow an object as it moves around a perimeter and zoom in on a particular scene for closer and clearer images. In most cases security operators control p/t/z cameras manually. Of course this means that operators need to first identify the potential threat and then move the camera to follow the perceived threat. This has several drawbacks, one being that it’s easy to lose track of a suspicious object when it gets out of the line of sight of a camera. The other drawback is that if an operator is manually guiding the camera, he can’t do much of anything else, like call for a response team.
Automatic tracking using p/t/z works by having fixed cameras (or other object detection sensors) identify a suspicious object and its location using advanced VA applications. The fixed camera then passes this information over to the p/t/z camera, which is able to track the object while automatically using its pan and tilt capabilities. Of no less importance is the p/t/z camera’s ability to zoom in on the tracked object to capture a more detailed image. This makes the p/t/z automatic tracking not only extremely helpful during real-time events, but also for identification purposes and after-the-fact investigation.
Multisite dashboards — As we know from living in a digital world, how information is displayed is just as important as what that information is, particularly so when you’re monitoring large amounts of sites, with many information streams. If this information is not displayed intuitively, something important can be easily missed. That’s where video surveillance dashboards can help.
Dashboards are particularly effective when you’re monitoring many sites from a central control room. Just imagine a railway company with dozens of train stations or a large retail bank with hundreds or thousands of branches, all with video surveillance cameras that need to be monitored.
The dashboard gives operators an at-a-glance enterprise-wide view of video system status and maintenance issues across every remote location, from one screen. This provides for easily identifying security, safety and maintenance problems, and responding according to their severity. Operators can also view and retrieve video of any location instantly simply by clicking on the designated location icon.
Page 1 of 2 pages 1
Video Surveillance ·
Video Analytics ·