Video Processing: A Full Analysis?
Video analytics has been around for many years, and much like VMD which preceded it, there are systems which perform very well, and there are those that are built to a price. Whilst it cannot be disputed that good quality analytics adds benefits to security solutions, should the industry be looking towards its role as a management tool?
Let’s start off by saying something that a few of the ‘marketeers’ within the industry will probably disagree with; it is a pointless exercise trying to quantify ‘Return on Investment’ for a security system. Many of those selling solutions will talk about how their system offers a better Return on Investment than their competitors. Evangelists for technologies will drone on about how their flavour of technology has a better Return on Investment than the technologies they are trying to disrupt. As the economic climate remains somewhat frosty, the Return on Investment arguments are getting louder. The problem is that where security is concerned, Return on Investment represents little more than guesswork.
Okay, now that a few hackles have been well and truly raised, let’s add to the argument by stating that every security system does in fact have a Return on Investment. Don’t be confused by this; the Return on Investment is most definitely there, but it’s very difficult to quantify exactly how great it is.
As an example, most security systems have a deterrent value. If an opportunist criminal passes a site by because the security in place deems it too great of a risk for the possible return, how can that be quantified? Is the end user going to be aware that the investment in security has done its job? Also, how can the value of potential losses from the crime that was prevented be qualified?
Obviously, the Return on Investment cannot be assessed with any certainty. This is something that those evaluating security systems will come up against, time and time again. A successful security system might only prove its worth to the end user on rare occasions, but that doesn’t mean it isn’t providing protection and getting results at other times.
Go beyond the ‘opportunist’ crook, and consider the career villain. These are the group of criminals that are responsible for the more significant losses which businesses might suffer. They will do their homework, and where credible and robust security is in place, a target might be deemed to high risk. The threat of losses will certainly be mitigated, but again, without some pie-in-the-sky accounting, it’s not possible to quantify a Return on Investment for that security system, even though it most certainly exists.
What is important here is to realise that this isn’t a pop at those who use the Return on Investment argument to sell. The purpose is to highlight that where a business or organisation is using Return on Investment as a gauge for future projects – and in today’s economic climate many are – justifying further investments for system upgrades or expansions can be a difficult task.
Things change dramatically, however, if the system in question offers additional benefits. They will change even faster if those benefits are realised on a day-to-day basis, and deliver tangible rewards for the organisation in question. Could the answer lie in using video analytics for management purposes?
Video analytics look for exceptions to expected behaviour, and unquestionably work best when specific details of the exception being watched for are known. One problem with security incidents is that often you are looking for the unexpected. Because criminals will always try to conceal what they are doing, the potential variables are manifold.
This creates challenges when configuring analytics, and goes some way to explaining why many who use the technology opt for the use of a sterile zone.
A sterile zone consists of some form of obvious demarkation between an area that people are permitted to be in, and an area which they clearly barred from. The idea of a sterile zone is not to physically restrict access to an area – although some sterile zones do just that – but to ensure that anyone entering the zone has done so with intent rather than accidentally. Typical approaches include fencing or other barriers, signage or audio-visual devices.
By creating a sterile zone, the idea is to prevent what could be innocuous or accidental activations. Any entry into the zone represents a violation, and this fact alone simplifies the creation of analytics rules, which in turn ensures that the performance of the video analytics engine is optimised.
This is worth remembering, because it is the simplicity of the sterile zone, the ease with which a rule can be created when any presence in the zone is an exception, that makes it so effective when compared to scenarios with a greater number of variables. If someone is in the sterile zone, then they either scaled a fence or disobeyed notices or forced an entry. Therefore, it is an alarm condition. All the analytics needs to do is look for a person in that area.
The biggest difference between security tasks and business management tasks is that typically, business management tasks are simpler to define. This is because the system is inevitably looking for something that is specified by the user, and is understood. Security use of video analytics typically looks for exceptions to normal behaviour, of which there can be many, some of which are very hard to predict. It is much simpler to consider behaviours that are expected, and which you want to measure or validate.
One example, which is used by a logistics company in Germany, is to use video analytics to track parcels in a delivery centre. A core function of the system is to monitor the automated system which sorts and manages parcels prior to loading for despatch.
Monitoring items on a conveyor belt is simple, because the potential exceptions are basic. If the belt stops running, or if an object is hanging off the side of the belt, or if an object appears or disappears on the belt, then an alert can be generated. These exceptions are all simple to define, and the analytics software will not need a complex set-up procedure to achieve this.
By eliminating the chance of random or unexpected exceptions, the resulting simplicity makes the video analytics more efficient, more reliable, and certainly more cost-effective as a result. It also delivers a very real benefit to the business, and in turn saves it money as any delays are immediayely notified, and items that go missing in the process can be located by going back to their last ‘appearance’ in the footage to see what has happened.
Video analytics technology can easily be utilised for a wide range of tasks beyond security, as is proven in this case. The technology can be used for site management tasks such as monitoring processes or production lines, for marketing and footfall analysis, for entrance control, people counting, vehicle management, etc.. Indeed, wherever a visual element – an action, event or identifier – is present, analytics can provide a solution. It matters not if the system is monitoring status LEDs or gauges on machinery, directing vehicles to empty loading bays, or even monitoring parcels on a conveyor belt, the technology will deliver benefits.
The retail sector is a good example of an industry that can gain significant benefits from the use of video analytics as a management tool. Whilst many in retail and reluctant to make further investments in security as the High Street suffers due to restrained consumer spending, it will invest in technologies that deliver business benefits. One very simple application for video analytics is queue management. The video surveillance system can flag up any queues at fitting rooms or Point of Sale locations in real time, enabling management to introduce extra staff or open new tills. This ensures that potential customers aren’t lost by simply walking away.
Out of working hours, video analytics can be scheduled to also provide a high degree of security as well. That a premises is closed in itself creates a sterile zone, thus allowing the technology to deliver quality results.
Video analytics undoubtedly offers much more than security. The technology can be used to instigate actions to a wide range of visual exceptions. That these can add value is a real bonus, added to by the fact that the technology also introduces the often quoted by rarely quantified Return on Investment.
End users will interact with their systems on a daily basis, and that interaction will be delivering management benefits, positive benefits, and that can increase their satisfaction and familiarity with the solutions they have invested in.
Many in the security industry don’t push the business benefits of analytics, but they should; as the potential on offer is significant in terms of opportunities.
|Video Analytics – a case in point
Benchmark recently met with the Facilities Manager for a large international retailer. One part of his remit was to provide security at numerous sites, and whilst his budget seemed healthy, a quick calculation of the number of locations that needed securing showed that corners would probably need to be cut. What further impacted on the situation was the fact that the business in question was self-insured.
Prior to specifying the solution which is currently in place (and being further rolled out), a number of systems were suggested by a range of installation and integration companies. All claimed to be able to offer a high degree of security for the allocated budget. However, all failed during proof-of-concept trials.
Eventually, a system that utilised video analytics was proposed. During out-of-hours periods, the analytics was used to protect the site and its assets. The video element was further boosted by using external detectors in outdoor areas, thereby allowing a ‘double-knock’ scenario to be created to minimise nuisance activations.
During working hours, the alarm and remote signalling element of the system was Unset. However, the analytics rules were altered by schedule, and were then utilised to assist management in the business.
The analytics allowed the marketing team to collect tracking and dwell data from numerous sites. This data could then be analysed and allowed behavioural patterns of customers to be built up.
Also, sales staff at the sites could access the system, with the analytics highlighting issues for queue management and sales staff deployment. Management were given tablets, which allowed them to access information and receive warnings from anywhere on site, and even when off-site.
The result was that the marketing and sales departments contributed to the system cost, so the budget was met. Also, satisfaction with the day-to-day management benefits played a significant role in the system being rolled out over all the business’ UK outlets.