Hi-Tech Security Solutions spoke to Protoclea’s CEO, Bennie Coetzer to find out what this local developer of security solutions is up to and what it intends to get up to in the coming year.
Hi-Tech Security Solutions: Protoclea has been in operation for some time. Tell us about the company.
Bennie Coetzer: Protoclea was formed in 2006 after Thales was sold to SAAB. We took the image processing business with the intention of building a company that can provide image processing technology to South Africa. While the SA Defence Force will always be a primary partner, we want to expand in the commercial business as well.
We obviously target the defence market, but also address the security and traffic markets. Our clients are high-end clients to whom security is important and would also be the target of sophisticated attacks, for which we provide world-leading solutions. In the traffic and vehicle management market we provide solutions ranging from very large parking management systems down to low-cost, single lane ANPR detecting (as in traffic offences) and access systems. Clients here would include owners of parking areas such as airports, shopping centres and even residential complexes.
HSS: Can you name a few projects you have been involved in?
Coetzer: Some clients do not wish to be named, but we have provided (with ZEAG) probably one of the largest and most successful parking management systems in the world to OR Tambo Airport, with success measured in the actual reduction (to zero) of vehicle theft as well as a substantial increase in revenue due to reducing fraudulent activities.
Other clients also include large corporations and the protection of national keypoints and of course the SANDF remains a major client and partner. We have also supplied solutions to provincial governments for traffic monitoring and so on.
HSS: Do you develop all your solutions locally? Do your customers see this as a value-added service?
Coetzer: Yes, we do. In addition to having a very capable development team we also have close collaboration with the CSIR with regards to algorithm development and we also support a number of research projects at local universities.
Being local means we can offer solutions that match local needs closely, we can develop specifically for specific needs if required, we can offer inexpensive training to consultants, installers and clients, and a host of other benefits, not to mention the rapid response and the quality of support.
As we concentrate on pro-active systems, the straightforward, reactive solutions offered by imported products (out of necessity due to the long-distance support link) simply cannot take into account local conditions. These do not only include local environmental conditions, but also our criminals’ behaviour, such as disregard for life, their brazen behaviour due to the high probability of getting away with it, our limited law-enforcement capabilities and the shortcomings of our judicial systems.
HSS: Many companies offer video analytics solutions and many people are sceptical about the effectiveness of analytics. What makes your offerings different?
Coetzer: The problem with video analytics is fundamentally a lack of understanding on the part of clients, installers and consultants. We have no doubt that it is the technology which, together with information processing, will form the most crucial aspects of security systems of the future. The lack of understanding comes from the fact that it is sold as a solution, whereas analytics is fundamentally an information reduction tool, ie, it takes in a huge amount of information and attempts to retrieve relevant parts of it. The failure to understand that analytics should, at least for now, be used in a support role rather than the expectation that it can be used in a decision-making role is what makes systems fail. Of course, the majority of systems are not able to use the analysis output in this way and hence the supposed poor performance.
Another limitation is that video analytics systems’ set up is by its very nature (especially outdoors) complex as it has to deal with a stochastic (random) environment. This requires that the analytics system has to provide the controls that can be appropriately adjusted, but also that the installer knows how to use them correctly.
And last, but not the least, all analytics systems are not the same. They all carry the key functionality, but do not execute the same algorithms to achieve results. With our analytics system, Nobilis, we perform continuous research and experimentation in reducing false alarms, and when used with our contextual analysis system we believe that even more reduced levels are possible, without lowering detection rates.
In addition to offering the generic solutions that anybody offers, we offer more functions than most as well as the ability to customise on the analytics level for specific problems, for example, whether miners wear the protected gear supplied, whether specific doors are closed before others are opened, confirming that where faces are to be recorded that the person did not turn his head away, weather detection to adjust performance parameters, high-speed vehicle type detection, and so on.
HSS: Do you work with specific hardware providers or are your systems open to whatever the client has in place?
Coetzer: In the markets that we operate in, quality and performance is more important than cost, although cost always plays a role. There are some providers we do prefer because of the quality of their products, but to be in business we have to be able to integrate with most. Where there is existing infrastructure we are happy to use it. If it does not provide the picture quality we need for video analytics, or the information for incident management, we inform the client about a drop in performance for further consideration.
Phorcas, our incident management system, has as one of its key functions an integration system. This means that we have to integrate to other systems and yes, openness and standards are important. We only provide image processing systems, we do not provide access control systems, cameras, etc, but we need them for a complete solution.
HSS: In your experience, have you found the expectations of customers differ from what technology is actually able to deliver?
Coetzer: Very much so, and the answer is both that they expected more and sometimes they expect less. One of the problems is that consultants are often not fully aware of video analytic possibilities and limitations. The salesperson obviously explains all the positives, but seldom the workarounds, the limitations, and so on. So visiting IFSEC may tell you about what products can claim, but more intense knowledge of the algorithms used will tell a consultant much more in how that kind of technology is applicable or not to a specific site’s needs.
Installers typically do not understand the security needs as a whole system – after all that is generally why the consultant was used. Making a system work and commissioning it is not the same as making the security work. For that an intimate knowledge of the threat is required, appropriate procedures specific to the site need to be developed and the installed systems have to be tweaked to support that, not a generic sign-off paper.
HSS: How do you see the company developing in the future?
Coetzer: We are firm believers in video analytics and information processing and these two form the basis of our new generation of products launched in October. With video analytics we will continue to improve algorithms and while this will be an incremental process we will get better and better at this. Emphasis is placed on object identification and tracking as well as behavioural traits. We want to see what you do, what you carry so that a threat can be assessed pro-actively.
When it comes to information processing we believe automation is key, simply because of the number of sensors that are available to a security system. Contextual analysis is therefore a new field which, we believe, will improve proactive prevention of crimes, and sophisticated crimes at that.
But we also have great faith in the mental abilities of humans. While we think they are often used incorrectly for tasks that are easily automated (and then of course are blamed for being ineffective), we have new features in our systems designed to get more out of humans. The future is not to use less human operators, but to use them better.
With respect to other systems such as Sinon, our number plate recognition system, we believe that detection can always be improved and will do so as algorithm research and camera technologies improve. The key feature here, however, is the ability to use the information. Thus ANPR is only part of it, the whole parking/access/traffic system is the full solution and that is where we are at.
© Technews Publishing (Pty) Ltd. | All Rights Reserved.
printer friendly version