Challenges of security surveillance….. See no evil

The best definition for physical security is available from the United States Geological Survey:

There is no object so well protected that it cannot be stolen, damaged, destroyed, or observed by unauthorized individuals. A balanced security system provides protection against a defined set of threats by informing the user of attempted intrusions and providing resistance to the would-be intruder’s attack paths (USGS, 2005, p. 12)

 The first part of this definition is common to all security efforts; we cannot stop a highly motivated attacker. It is in the second sentence of our definition that we find the objectives of physical security. Said another way, the purpose of physical security is to delay an intruder’s advance toward a target long enough to detect and respond with human intervention. Human intervention includes on-site security officers, armed response, police, or other relevant human controls.

 Achieving physical security objectives requires policies, standards, guidelines, and controls addressing prevention, detection, delay, response, and assessment

With the ever-increasing risk and escalation in the frequency and severity of the crime and confrontational violence on businesses, hospitality, fast-casual dining, education, medical, residential sectors of the spectrum — students, guests, patrons, tenants, residences, staff and visitors will avoid places that have been victims of crime and violence for fear of their own life safety which will result in the decreasing of their willingness to frequent, purchase for fear of the own life safety which will ultimately have a direct negative impact on one’s business profits and brand reputation.

Regrettably, business guarding mentality crime response approach by beefing up and throwing monies for costly additional security guards, armed guards and or “tactical response” in soft skin vehicles from the 9,000+ security guarding companies, is not only adding significant cost to already strained budgets but regrettably still exposes and leaves room for human error and often after the fact slow first response critical incident management times resulting in devastating losses/brand reputation/duty of care failure 

Whilst there have been huge positive inroads around LPR, Facial Recognition, COVID Temperature, audio, Intrusion and Perimeter detection and the adoption of fight back technology (smoke cloak etc) many businesses, hotels, medical centres, schools, residential estates across South Africa continue to be exposed to risk despite having a myriad of cctv cameras that are either:

• Not monitored and only used reactively when reviewing a post tipping point after an incident occurs or is discovered, which defeats the objective of risk prevention/mitigation. In addition with all the hundreds of cameras installed, running 24/7, accumulating an enormous amount of data thus making data extraction quite a challenge when reviewing a post-incident which can also be extremely difficult and time-consuming.
• Reactive surveillance (response to a specific incident, either retrospectively or in real-time) via localised monitoring on-site in most instances by Grade C Security Guards (for cost reasons due to price-competitive reasons) by means of large banks of video wall monitors which fail to instantly identify threats as many of the unintelligent devices lack the ability to effectively detect objects such as guns, knives, violence or common behaviours leading up to the incident. This is further negatively compounded by unintelligent motion detection cameras that are highly erroneous, triggering against everything that moves, be it insects, an animal or even a shadow change, resulting in false alarms being triggered which can easily exceed 100+ false triggers per camera per day.
• Combination of reactive and proactive surveillance (via A) via remotely monitored off-site in most instances by Grade C Security Guards (for cost reasons due to price-competitive pressures) by means of large banks of video wall monitors which fail to instantly identify threats as many of the unintelligent devices lack the ability to effectively detect objects such as guns, knives, violence or common behaviours leading up to the incident. This is further negatively compounded by unintelligent motion detection cameras that are highly erroneous, triggering against everything that moves, be it insects, an animal, wind or even a shadow change, resulting in false alarms being triggered which can easily exceed 100+ false triggers per camera per day.
• Moreover, the number of potential camera feeds to select from often greatly exceeds the number of available video wall monitors on which to view them where the impossible task is often given to a single operator often responsible for up to 175 cameras at one time

High expectations are placed upon surveillance technologies to protect infrastructure and public places and optimal design and development of new technology should not focus solely on the capabilities of the system itself, but on supporting the operator’s cognitive vulnerabilities. 

Further to this, I am still left at odds by end-users camera criteria selection being solely around cost and only cost and the cheaper the better.

Very little consideration is placed on the camera targeted observation matrix levels of differentiation required to ensure effective surveillance:

Detection – camera with 1.5 pixels will detect an unidentifiable object.

Classification- camera with 6 pixels analytics will be able to distinguish between an animate (animal/person) and or an inanimate (vehicle object

Recognition - camera with 12 pixels where an object can be distinguished as specifically as a person

Identification - camera with 25 pixels where the identity of a person is easily distinguishable

At present, the biggest challenge faced by businesses hospitality, fast-casual dining, education, medical, residential sectors of the spectrum is to do more with fewer resources.

Security enterprise virtual surveillance architecture objective and design high-level security principles which are set out using the following considerations:

• Deter – stop or displace the attack
• Detect – verify an attack, initiate the response
• Delay – prevent the attack from reaching the asset
• Respond – apprehend the attack and prevent further progress of the attack
• Mitigate – minimise the consequences of an attack
• Mitigate the need for having to keep on adding new hardware for SOC&CC monitoring
• Reduce the manned guarding component in lieu of adopting leading-edge disruptive technology and AI
• Virtual visual dashboard business Intelligence reporting from all integrated echo systems
• Central storage vault for all alarms as well as “live video vetting”
• Central audit trail to track and measure SOC&CC operators responses and performance
• Central virtual CEM, digital SOP’s for all sites
• Scalable
• Enterprise open
• Leverage cloud computing and augment one's surveillance systems and move away from on-premise DVR / NVR hardware and dedicated PC for storing video files.
• Overcome Internet connection issues
• Flexible vertical software architecture allows centralized processing, decentralized edge processing, or the combination of them, making security system upgrades possible for those customers who have limitations in their local infrastructure.
• Timely identification and notification to site security personnel,
• Ability to digitally or optically zoom in to the area where intrusion was detected
• Ability to via speakers communicate and alert the “intruders” their presence has been detected.
• Ability to turn on lights, sound a siren, issue QR codes, OTP, open doors and trigger fight back technology
• Negligible false alarms being triggered,
• Provide real-time virtual situational awareness, compliance, prevention, identification and detection of risks across one entire business operation.

These principles should be used to direct the thinking around potential security measures required to mitigate each of the identified risks

These surveillance objectives are all critical requirements needed to identify and detect real-time risks, so that one can make informed, accurate and immediate first response decisions to whatever the situation may be and not have to worry about human error where costly security guards might have deserted site, are drunk, are sleeping on duty, have been compromised or a new guard that has never worked on the site has been dropped with zero site training and or orientation etc or allow any slow poor and ineffective response times.

This is especially true when one considers the violence and severity of criminal activity clock exposure are at its highest after hours and or over weekends where businesses have left the management have often left their business the hands of poorly paid and or poorly capacitated security guards, along with the none monitoring of businesses cameras etc.

 Businesses, hospitality, fast-casual dining, education, medical, residential sectors need to consider how to enrich, augment and sweat their existing surveillance platforms and look at developing an enterprise technology strategy to address risks identified that will to provide

• Data Analysis were real-time virtual AI threat and detection surveillance technology will open up a statistics dashboard to provide event triggers and notification of the threat in order activate one pre-defined priority tiered CEM and first response
• Instantly notify CEM team and external agencies distributed and escalated through proper security and web channels of the incident
• Reduce negligible false positives to ensure effective detection and first response time is faster
• Powerful identification and detection system is the first line of defences in-depth and is most effective when the detection response time is faster and there are negligible false positives.
• Constantly analyze video streams for anomaly detection, which tracks and reports in real-time including sending pre-defined tier risk CEM escalation notifications and alerts.
• Increase one’s speed and efficiency in the identification of threat detections, ensuring critical management of response timelines, enable activation of first responder and external agencies effective response on scene. This real-time information allows one to create an informed and appropriate first response to whatever the CEM situation may be.
• Other surveillance echo system integration includes:
• Live body camera remote monitoring with audio
• Drones
• Remote real-time covert monitoring with audio
• Vehicle surveillance, LPR and audio
• Remote mobile and or fixed temporary site/location monitoring
• Mobile forward command – JOC / VOC

There needs to be a massive change — one that can only be brought about by deploying cutting-edge fast free false alarm virtual disruptive technologies and AI designed to provide real-time situational awareness across multiple sites by detecting behavioural pattern recognition, threat anomaly detection, as well as equip first responders responding to potentially dangerous situations in an informed manner.

Real-time threat detection and loss prevention are needed to identify behavioural patterns, detect suspects, and even escalate and report threats to the concerned CEM team and external agencies.

This is all possible

Chris Cobb

KIFARU Risk Management

Email:chris@kifarumanagement.co.za

AGILE-PROTECT-DEFEND