AI-powered video analysis for Door Forced Open scenarios

If you use the words "Door Forced Open" when talking about Building Security with anyone who’s not in Physical Security, they’ll probably say something like “Wow, is that like someone using a crowbar to force the door and gain entry?” I must admit, the first time I heard the term, I thought it was a pretty bad situation.

However, if you use the term (or it’s acronym – DFO) with anyone with a Physical Security background, they’ll probably just roll their eyes at you.

What I quickly learned from talking to my customers, was that these DFO events were not the occasional but important events that I imagined them to be, but were instead the bane of their lives, and causing an inordinate amount of noise, with each individual alarm having to be checked & verified, but with over 90% of them being a false alarm!

I got some security insiders to explain how this could be the case and learned that most Physical Access Control Systems are reliant on what turns out to be pretty ‘dumb’ sensors in the doors and that it’s these sensors that cause the alarm to be raised. In trying to understand what causes these false alarms, there were 2 situations that I heard regularly

• Daily Cleaning: A cleaner will be cleaning the glass on the door and the pressure of them wiping the glass from the inside is enough to take the sensor out of alignment and cause a DFO alarm to be raised.
• The Lunch / Evening Rush: I’ve probably been the cause of this one myself. In the excitement to leave the building, people leaving through the door may push on the door, before the handle or push bar is fully open, also taking the sensor out of alignment and raising another DFO alarm.

What I also quickly learned is that not only are there a bunch of these being fired every day, but after getting the alarm, moving over to the camera system, finding the camera that’s watching that door, going back to the alarm time and viewing what happened — that it could take upwards of 4 minutes to verify if this alarm was noise or something that needed to be dispatched to.

How Ambient.AI is helping.

Over the last few months, we’ve connected an integration to some forward-looking customer’s PACS systems. By having direct access to this data coming in from the readers and sensors and mapping those doors to the cameras we are proactively monitoring for them, we’ve seen some huge potential benefits.

• By being able to monitor the door as the DFO alarm is triggered, we can see if anyone actually entered the building. If they did – we would raise that as an alert to be reviewed. If people only exited the building (or in the case of the daily cleaning), did not exit at all, we would not raise an alert, because there was no security threat involved. This significantly reduces the noise our customers see with DFO Alarms today by upwards of 85%.
• With our camera mapping and monitoring, the alert that Ambient would raise would include the PACS info as well as the camera footage at that exact moment when the DFO was triggered and we saw someone enter the building, significantly reducing the time and effort an Operator takes to review the alert and decide on a course of action, taking the average review time from 4 minutes to under 30 seconds.
• If required, the Operator can quickly dispatch a Patrol Guard to attend, directly from within Ambient. The patrol guard receives the Dispatch Request to their phone and it includes a clip of the incident, so they have an exact visual description of who they are looking for once they arrive.

We’re now entering the next exciting phase with these customers starting to turn off their DFO alarms in certain areas that Ambient is monitoring so that they can start benefiting from we’ve already proved in the Pilot — our ability to reduce 100’s or 1000’s of those alarms from being raised as well as allowing Operators to take much quicker action on those that are raised.