Many end user organizations attach great importance to perimeter security, which is known to create certain issues, for example false alarms. In this regard, AI can come in to help.
Many end user organizations attach great importance to
perimeter security, which is known to create certain issues, for example false alarms. In this regard, AI can come in to help.
For many end user organizations, protecting the perimeter is a top priority. Yet perimeter security, while important, is known to cause certain problems for users. One of them is false alarms – mistaking animals or wavering grass for intruders, for example. Meanwhile, end users also demand certain features from their perimeter security solutions, for example smart search and business intelligence capabilities.
[See also: How to deal with the single biggest challenge in perimeter security]
In this regard,
AI can come in handy. “Using AI to identify and verify suspicious activity can yield more accurate results and reduce the number of false positives and therefore the number of false alarms – something which those monitoring large perimeters are keen to avoid. AI-enabled cameras can filter out benign activity, having been trained for particular use cases such as identifying people and vehicles. Also, customers now receive cameras with these calibrations configured out of the box, enabling the end user or system integrator to easily set up the solution,” said Niklas Rosell, Global Product Manager for Smart Camera Features at Axis Communications.
Benefits of using AI in perimeter security
Perimeter security is all about detecting intrusion, loitering and other suspicious activities along the perimeter. Any AI-based analytics that can help users meet those objectives can be beneficial. “The most common AI-based video analytics found in perimeter security are: outdoor-optimized people and vehicle tracking, which generates far fewer false alarms than simple motion or pixel-change detection; automatic license plate recognition (ALPR), which streamlines operations and record generation/reporting; and face recognition, which can be used as part of 2-factor authentication systems for high-security facilities. Other AI-based analytics also exist, of course, but these typically serve niche or facility-specific requirements and are therefore less common in perimeter security applications,” said Brad Martin, Director of Product Management at Senstar.
While AI in perimeter security is often used with video, it can also be applied to data from other sensors. “AI can be applied to any complex signal and be used to identify key patterns, typically obtained from deep-learning and other AI techniques. However, mission critical systems, such as those found in high-security applications, need to handle unseen data or new situations in a predicable, controlled manner. This is where experience and careful system design comes into play. AI-enhanced systems can achieve higher levels of performance in areas like nuisance/false alarm reduction but must include finely tuned fallback mechanisms to guarantee detection in all valid but potentially untested scenarios,” Martin said.
That said, below are some of the benefits of using AI in perimeter security.
Reducing false alarms
As mentioned, one of the biggest challenges in perimeter security is false alarms, which can be addressed by AI. “For perimeter security applications, enhancing intrusion sensors with AI technologies has the potential to defeat false or nuisance alarms while maximizing the probability of detection (PD). For instance, consider a perimeter adjacent to a high-traffic public area. AI technologies can distinguish the presence of people or vehicles near the perimeter from small animals, debris or environmental effects, thus eliminating a common source of nuisance alarms, while the fence sensor continues to detect actual disturbances on the fence, avoiding the potential for false alarms generated by innocuous human activity near the perimeter, such as simply standing or walking along a public-facing fence,” Martin said.
“The two main challenges of any perimeter security solutions will always be ensuring a high probability of detection whilst achieving a very low nuisance alarm rate. Both are needed to retain operator confidence,” said Mark Horton, VP of Bandweaver Technology. “Bandweaver utilizes AI in our perimeter security solutions to filter out environmental noise caused by traffic, weather, etc. whilst still detecting any intruder. We deploy the latest generation algorithms, DNN, RNN and other machine learning techniques to achieve market leading detection with minimal nuisance alarms.”
Smart search
AI also accelerates
smart search, which is critical during investigation. “The ability to quickly search large volumes of video for specific patterns, such as a specific face, a color of vehicle, or even a hat, has the potential for both massive cost savings with respect to operator resources as well as identify events that humans might simply miss. In addition, the reduction of nuisance and false alarms will also assist organizations by reducing the overall volume of clearly irrelevant events,” Martin said.
Cost savings
When engineered correctly, AI solutions provide significant operational cost savings. “At a high level, the automation of error-prone tasks saves the cost of sending officers to investigate and resolve incidents – in this way AI is a force multiplier. Indirect cost savings include lower insurance costs, and compliance costs in addition to the direct costs sue to loss or liability brought on by intrusions. However to realize these significant cost-saving intrusion prevention technologies need to be accurate with low false positives,” said Srinath Kalluri, CEO of Oyla.
Business intelligence
Finally, AI and provide a wealth of insights for users who can better plan for their perimeter security. “From a business intelligence perspective, high quality data analytics is absolutely crucial and AI can perform these processes quickly and at scale. This will enable businesses to gain a better understanding of the threats commonly seen at the perimeter, plan security strategies to protect the premises and implement the right solutions to support,” Rosell said.
“AI technology is data driven. The key goal of video based perimeter AI technology is to glean insights from this data to better emulate human functions for real time security operations. However, the same data can be used over longer periods to gain insights and trends that help security professionals optimize operations. In the
retail space, AI is already being used to understand patterns of movement, traffic and purchase influencing the design of space to best enhance customer experience. Similarly in the perimeter context, security professions are starting to use AI to design checkpoints, ingress and egress points, physical barriers etc., using AI gleaned incident data,” Kalluri said.