What are the CID’s plans for camera monitoring and how exactly does camera monitoring work?

The exact model that the CID will use will depend on negotiations with service providers closer to the time of implementation, but the following is a guide to how camera monitoring typically works:

Modern cameras use intelligent analytics to trigger an alert when a human figure, and/or vehicle is detected, depending on how the analytics are set. This technology is improving all the time, and forms the first layer of filtering. 

An example of how this integrates with a public safety and security system is as follows:

Professional monitoring companies typically employ what is known as ‘black screen monitoring’, whereby the cameras generate an alert once it has seen either a person or vehicle or both (depending on settings). This alert is sent through to the monitoring company in the form of an image or short clip which goes into the monitoring company’s queue system. 

This is very different to old systems, which required hundreds of screens to be monitored by a small number of individuals who have to watch live feeds constantly. (Have you ever tried to watch hundreds of movies at the same time?)

At the monitoring company, typically a team of human operators are waiting for the alerts to come in and just like at a bank, an automated system ensures that the first operator free takes the oldest alert.

The operator will view and verify the alert, often within seconds. They have the ability to view the image or clip generated by the alert which is often a clip starting a few seconds before the alert on the camera to a few seconds after the alert. They also have the ability to view that camera, and any connected cameras, live.

Different environments can have different protocols assigned. For example, an area where people shouldn’t be walking at night might be set to escalate if any person is seen, whereas areas where people are known to move through regularly may be set to alert only on suspicious activity.

The human operator is trained to make a judgement as to whether the activity that triggered the alert amounts to suspicious behaviour. If it is not considered suspicious, no action is taken. If it is considered suspicious, the human operator then escalates the alert as per the protocol which could be to a more specialised control room, or directly to an armed response company.

This forms the second layer of filtering.

A more specialised control room comprises a core group of experienced controllers who work closely with each other and the various response resources. For this, they require direct and intimate knowledge of every street and green belt area, as well as an understanding of the community, and the resources available to the community.

What this means is that when an alert is escalated to them by the monitoring company, they are able to dispatch a physical resource to intercept while they digitally investigate. With a thorough knowledge of the area and past activity, they can predict movement of the individual(s), open up relevant cameras close by and monitor the individual(s) on live view, or open cameras in areas where they predict the individual(s) may show up. 

They can perform playback of the alert to see what the monitoring company considered suspicious and use their knowledge of the community to guide their reaction. For example, they may know that the person seen is a particular resident that walks their dog every evening at 22:00 or a worker that arrives in the area at 04:00am and requires no further action. 

They know through the feedback and relationships with the community’s response resources who the trouble makers are and who the individuals are that are unlikely to be a threat. Therefore they know if they need to send a response, or just monitor the individual(s) to know they leave the area and when in doubt, protocol dictates that a response is sent to investigate further. 

If necessary, the response will be coordinated between different service providers (including, for example, private alarm monitoring response companies).

This specialised control room, which takes alerts escalated from a professional monitoring company and uses intimate local knowledge to inform the correct response, is the model that CBCSI currently uses for street scheme cameras.

This forms the third layer of filtering.

In the case of the proposed CID, the response to any camera alerts would come from one the CID’s three tactical vehicles in the first instance, co-ordinated by a control room and the CID’s Public Safety Officer.

With analytic cameras, the traffic flow (human or vehicle) dictates the daily start and end timing for camera monitoring. Usually in a high traffic area, monitoring is only active during night time hours to avoid overloading the control room with alerts.

LPR (Licence Plate Recognition) cameras will also be used by the CID. This is where an alert is generated by the number plate of a flagged vehicle as dictated by the database in use. The database in use for the CID would be one shared by law enforcement agencies City-wide. 

In addition, all cameras footage can be used for evidentiary purposes with video footage being saved and handed to the relevant authority.

A Standard Operating Procedure is agreed by all parties in advance to determine exactly how alerts are handled. Typically, detailed reports are provided monthly on the number of alerts, frequency and timing of alerts for each camera, whether a response was considered necessary, the time that elapses between each stage of the process, and the final outcome. Full audit trails are available. 

If a Response (eg Tac vehicle) is sent to investigate an alert, what can they actually do in practice?

In the absence of evidence of wrong-doing, there is not very much a response vehicle can do except politely question the person, as everyone has the right to movement in public. If the person was intent on committing a crime however, they will have noted that their movements were detected by camera analytics, and that a response was forthcoming. Experience indicates that this alone is a strong deterrent and the suspect will likely move on to an area where they can operate with a smaller probability of detection.

If the cameras can show that a person is behaving in a way that indicates intent to commit a crime, the cameras allow for response to be positioned close by and to monitor them and if a crime is committed, an arrest can be made.

If a responding officer has evidence from a camera alert that a Schedule 1 crime (serious crime) has been committed, then they can and will engage with a suspect, and make a citizens arrest if required.

Under the proposals for a CID, a dedicated Law Enforcement Officer would be present in a tactical vehicle at all times. These officers have full powers to stop, search,  arrest and fine, thus enabling a much harder-hitting response than we can achieve currently with only security officers. 

It is also important to note that interactions between armed response and suspects (including questioning) generate information and intelligence. For example, knowing that someone was in an area around a particular time and knowing that an incident occurred before that or thereafter is intelligence that can be used to show patterns that point to individuals which can, and do, lead to breakthroughs in investigations.

It is the combination of all the above that makes monitored cameras a crucial part of the CID strategy for Safety and Security. It is also crucial that all key areas are covered by the cameras. If they are not, experience shows that the crime migrates to the areas that are not covered, thereby shifting the problem rather than solving it. With all the key areas of Clifton and Camps Bay covered by a CID, the problems will be shifted out of our neighbourhood entirely.