AI technology can enhance operational activities and save central stations money, says Chris Brown, CEO, Immix.
In many of today’s central stations, video is an essential piece of the business.
Whether that monitoring provider offers a full-scale remote video guarding operation or, if they simply offer basic video verification of intrusion alarms, video plays a role.
While video most certainly improves the quality and reliability of service delivered by a central station, it can often create considerable added labor, cost and increases in operator response times if not used “intelligently”.
Simple motion detection, particularly in outdoor applications, is extremely unreliable as a detection measure for a central station.
It can lead to a massive amount of false alarm traffic that operators must process, creating an out-of-control alarm queue that can put every customer that center monitors at risk.
When providing video verification, if the central station is not using an automation platform that can accurately associate specific cameras with specific alarm points and launch those cameras immediately upon the receipt of an alarm, it can take an operator more time to sift through live cameras to find any video that is relevant to either initiate a response or dismiss an event.
The question then becomes, “How can we ensure a way for a central station to reliably receive video data that is relevant to them while, at the same time, preventing irrelevant video data from needing to be processed by an operator or agent?”
Video analytics was the first solution to truly address the need for intelligent video.
Starting roughly 20 years ago, commercial video manufacturers began developing technology to identify specific objects (person, vehicle) or patterns of movement (direction, line crossing) that could be considered security threats with the intent on getting only relevant video information to a user so they could take whatever action was deemed necessary in response.
While it took several years to really catch on, this technology made it possible for traditional central stations to begin incorporating more video applications into their operations.
Early video analytic solutions were typically hardware systems with embedded software algorithms.
Reliability varied from manufacturer to manufacturer, as did the cost of different products.
It was often quite expensive to put in place the necessary hardware or additional components required to achieve a reliable and efficient “smart” video solution, limiting the type and number of companies that were willing to build a business model around managed and monitored video services.
Over time, consolidation among video analytic companies ensued, further limiting the ability for many companies to adequately compete using the video analytics hardware/software combination business model.
This development led to the rise of the AI model as applied to video.
This model enabled smaller, more robust companies to develop intelligent algorithms that were essentially “hardware agnostic”, meaning they could be applied to any camera regardless of manufacturer whether via a video management system (VMS), firmware update or cloud-hosted service.
This has created a new method for central stations to easily apply AI verification to more cameras, thus improving their cost and labor efficiencies and, as a result, greatly enhance their overall monitoring operation.
AI verificationis a process that utilizes an AI video solution to review video alarm clips triggered by any form of video detection or other alarming solution that includes video footage and applies an AI “filter” to verify that the threat detected is real and actionable based on a preset threshold.
If those necessary criteria are met, then and only then is that alarm and associated video presented to a central station operator/agent to process and act upon.
AI verification is designed to reduce the number of false or nuisance alarms/events that a central station operator/agent must handle, allowing the station to increase the overall number of sites and/or customers they can support without having to increase the number of staff or add infrastructure, thus improving margin and efficiency across the operation.
A more recent use of AI technology within the central station is in relation to video or virtual guard patrols.
Many monitoring centers perform scheduled patrols using the existing cameras at a given site to replace or augment the need to have a physical security guard perform an in-person patrol.
Virtual patrols are much more efficient as they can be done faster, more frequently and cover more area in a smaller window of time than a physical patrol.
Not to mention, the cost to conduct virtual patrols is much lower than physical patrols.
Monitoring stations can now improve on those cost and operational efficiencies using AI technology.
Until recently, video patrols still needed to be conducted manually by a station operator – whether on a scheduled basis, or randomly throughout the day.
With the progression of AI, stations can now enable these patrols to truly be conducted “virtually” using AI.
A patrol can be scheduled and conducted behind the scenes with the AI programmed to look for certain anomalies such as scene change, obstructed views, device offline, person or moving vehicle detected, fire/smoke detection, crowd detection/queue management, fighting/aggressive behavior, littering/dumping, package detection and more.
In the event that an anomaly is detected, the AI will raise the patrol to a human operator to determine the next action to be taken.
This can dramatically improve labor and time efficiencies in a monitoring center, allowing operators to focus on critical threats and core competencies and thus, create additional cost efficiencies.
Reduced alarm traffic and labor time means a station can either reduce manpower or renew existing manpower focus on other core competencies.
This creates increased operational efficiencies which can then allow centers to offer new services that previously may have been unattainable due to labor constraints associated with false or nuisance alarm volume.
New services obviously lead to previously unrealized revenue streams, thus maximizing cost efficiencies and overall business growth and value.
For example, by incorporating AI into its video operation, perhaps a center may now be able to offer managed ingress/egress services for gate or managed door environments.
Such applications are typically considered premium services, and therefore carry a premium price tag, dramatically enhancing the value of your service offering.
To deliver a solution that will change the game for a central station, it all starts with using the best AI video technology available.
For a central station software automation provider to offer a true AI verification solution, it must first identify the absolute best-of-breed AI video manufacturers and partner with them.
It is then important to embed the AI functionality directly within the platform architecture so that it streamlines the business model.
This allows users to select specific sites, cameras and alarms and have all traffic from those points “scrubbed” against a pre-determined “percentage of confidence” threshold before it ever hits an alarm queue, to determine if there is a viable event that needs to be presented to an operator for processing.
Such a solution has been proven to reduce video alarm traffic by as much as 85%.
Finally, you need to determine if your business would be more amenable to a cloud-based solution or an on-premises solution.
Both are available and there are infrastructure and liability concerns that need to be heavily weighed in determining which option is right for your monitoring operation.
The technology is finally available that can truly make a difference for central stations looking to offer a robust, effective and highly scalable video monitoring operation.
This article was originally published in the August edition of Security Journal Americas. To read your FREE digital edition, click here.