Intelligent surveillance turns away intruders
When you manage a facility or enterprise, you understand the wide range of safety, security and liability risks occurring across the property. You’ve conducted a site assessment and survey and are doing what you can to create layers of protection from outside-in, targeting potential vulnerabilities.
The reality is that property crimes continue to track upwards, plaguing outdoor assets at commercial properties, retail businesses and educational campuses. Persistent loitering, vandalism, dumpster-diving, break-ins and theft perplex even the most seasoned facility executive, creating liabilities and unsafe environments.
For those charged with managing vehicle fleets, the challenges are compounded. Auto theft has risen dramatically as thieves steal vehicles or components, like catalytic converters and even fuel. Some 745,000 vehicles were stolen during the first three quarters of 2022, according to the National Insurance Crime Bureau (NICB), equating to more than $6 billion in losses and an increase of 24% compared to the same period in 2019.
AI revolution in surveillance
Broad application of Artificial Intelligence (AI) in camera-based security is leading a revolution in automation of intrusion detection followed by immediate deterrence of unwanted intrusions – elevating the effectiveness of camera-based security. Surveillance cameras have become active sensors, equipped with automated controls and AI algorithms that accurately detect intruders and alert before an actual theft occurs and without the delays of human intervention. Culled data combined with machine learning continually tailors to the environment, so the latest algorithms yield superior detection and appropriate responses to intrusion.
As an example of how the product category of video is advancing, a Silicon Valley manufacturer has developed smart automated surveillance guardians using AI that automatically detects and deters intruders within seconds, reporting directly to management or guards via smartphone or iPad. The automated deterrence has proven to be over 98% effective in moving intruders off the property before any damage is incurred. These smart guardians are easy to install because they directly replace building floodlights on a property and can also be connected through Power over Ethernet (PoE) to existing cabling infrastructures.
The smart guardian nodes are equipped to detect an intruder with 10,000 lumens of LED lighting, a 4K high-resolution camera, quad-core computer continuously performing image analytics, motion detectors, a microphone and integrated AI. They automatically deter unwanted intrusions with red/blue strobing LEDs, three-way loudspeakers for voice-down messages and ear-piercing alarms.
Upon detection of an intrusion, the guardian nodes send a notification with the event recording clip to the security personnel or facility manager. Authorized personnel can then assess if the automated deterrence was successful and if not, engage the intruder directly with push-to-talk messages or increase the automatic deterrence actions built into the guardians. These types of surveillance systems can be used as standalone solutions – with the ability to integrate with centralized video monitoring services to quickly dispatch authorities if necessary.
Convolutional Neural Network (CNN) algorithms within the AI application recognize intrusion and take immediate and fully automated deterrence actions that surprise intruders and defy predictability – yielding nearly 100% success rate in deterring unwanted intrusion. In deep learning, CNN is a class of artificial neural network (ANN), most commonly applied to analyze visual imagery. CNN analyzes images and video streams, offering stellar performance in speech and audio inputs. Proper application of CNN also determines the success of the automated deterrence, thus focusing human operators on real threats and safety concerns that require further intervention.
AI on a roll
AI in surveillance has changed dramatically since its early stages as simple motion-based detection. In its beginnings, algorithms would compare pixel changes from scenes in moving video to assess the detection area. Innovation in deep neural networks has also improved the technology with layers of data which resolve problems or shortcomings in computer vision. Deep neural networks make objects and their specific attributes detectable with high degrees of accuracy.
Video surveillance advancements have made cameras a cost-effective way to bring immediate detection and deterrence. It’s turned the product category into a proactive tool that can be leveraged for safety, security and operations. Installing cameras as a reactionary measure to only record and playback video events after the fact isn’t the best use of cameras, given the cost and labor for live monitoring. Instead, AI and analytics improve the efficiency of monitoring intrusion around the premises.
Real Time Streaming Protocols (RTSP) available in intelligent cameras easily integrate video streams into onsite vendor Network Video Recorders (NVR) and Video Management Systems (VMS). Facility managers can replace existing, passive IP cameras by reusing Ethernet cable and integrating the RTSP video to a current NVR or VMS.
What challenges do you face daily? Do you manage parking structures or an expansive, multi-acre site? A hybrid approach to security, using guard services for critical areas and coupling it with technology like video with license plate recognition and other intrusion detection could be the answer. Video can augment human personnel and guard services, helping offset current hiring and labor shortages while adding the benefit of electronic technology for early warning—giving personnel more information for a better, more efficient response.
More value in video
Fully automated and intelligent video deterrence and detection can be flexibly configured to meet end-user needs and business requirements. For example, it can suspend all monitoring activities during normal business hours or continue to simply monitor and record pedestrian and vehicle traffic. After business hours or late in the evening, the system can welcome guests or customers with a soft message accompanied by a slight increase in the LED illumination for safety. As the night progresses, the system can be used to inform front desk staff that a customer has arrived, pop-up a live video feed and activate a series of LED lights to establish a direct and safe path to the front entrance. The immediate reception is welcoming to guests, whereas intruders, looking to break into a vehicle or commence other criminal activity, will know they have been noticed.
The old way of deploying and watching hundreds and thousands of cameras for suspicious behavior and then reacting is over. The new way, which is far more effective, is deploying intelligent surveillance solutions with AI and analytics – turning cameras into real-time detection sensors.
The article appeared in Facility Executive, April 2023