AI solution saving end-customer €7200/year

“We have a client with perimeter security, that uses sensor-based alarms we have installed using surveillance cameras”, explains Mantas. “Due to the specific location, this territory cannot be completely gated. This means that both people with good and potentially bad intentions occasionally cross the perimeter, setting off a motion based sensor alarm.”

An abundance of alarms without any clear reason

“Before implementing IRIS™, our operators would connect to the installed video system remotely after receiving an alarm, trying to identify who or what had caused the alarm. However, in most cases, our operators were left baffled, not being able to find a perpetrator. There were relatively many of these alarms too – usually at least two each night, adding up to around 40-60 alarms per month.”

Short, non-threatening visits causing an excess of administration

“After some research, we found out that most of these alarms were triggered by scheduled postmen arriving during different times of the night. They were spending only about a minute in the territory, so our SOC operators had never seen them live. They were arriving at similar, but not exactly the same time every night. So in each case, there was always a possibility that there was someone else trying to get into the territory and not the postmen – meaning that each time the alarm must be considered a real threat.”

Limited options

“At that time, we had only two options to deal with these alarms:

1. Check all video records leading up to the alarm in detail to try and identify what had triggered the event, or
2. Sending a patrol every time an alarm was triggered without any apparent reason, just to be on the safe side”.

Costly and inefficient

“Both these options were problematic, time consuming and inefficient. Looking through the records would take the operators at least 5 times longer than checking the live stream. Our estimation was that this work would cost us at least €80/month. The second option was to not go through the records, but instead send a patrol to check the situation on site. The cost of this was at least €600/month and despite being a more reliable security assessment, the increased cost was of course not acceptable for the customer. In both cases it was a frustrating experience of spending a lot of time searching, knowing that you will probably just see the same postman again.”

The solution

“So we decided to choose the third option instead – connecting the cameras to IRIS™ analytics. This would allow operators to quickly see human movement in the territory and identify the good postmen as they arrive. Using the IRIS™ system we get instant notifications of movement in the territory and can therefore quickly identify persons without having to go through video records. In cases when there is someone suspicious in the area, the operators can quickly dispatch our patrol team to the site for additional verification.

After the introduction of IRIS, we have noticed that the number of incidents when alarms were not correctly handled by the video surveillance operator has decreased significantly. From 1-2 incidents per month to 3 incidents in the last year, over all the sites that we have implemented IRIS on.

The cost of this service is lower than the work cost for our operators in both methods described above. Additionally, the way it works now is much more reliable, excludes many false alarms and thus improves the morale of the patrols and our SOC operators.”

– Mantas Dirgėla, Guarding and Technical Surveillance Manager, G4S Lithuania