Artificial Intelligence (AI) trends in physical security systems

Download PDF version

If you’ve been paying attention over the last twelve months, you will have noticed that deep learning techniques and artificial intelligence (AI) are making waves in the physical security market, with manufacturers eagerly adopting these buzzwords at the industry's biggest trade shows. With all the hype, security professionals are curious to know what these terms really mean, and how these technologies can boost real-world security system performance.

The growing number of applications of deep learning technology and AI in physical security is a clear indication that these are more than a passing fad. This review of some of our most comprehensive articles on these topics shows that AI is an all-pervasive trend that the physical security industry will do well to embrace quickly. Here, we examine the opportunities that artificial intelligence presents for smart security applications, and look back at how some of the leading security companies are adapting to respond to rapidly-changing expectations:

What is deep learning technology?

Machine Learning involves collecting large amounts of data related to a problem, training a model using this data and employing this model to process new data. Recently, there have been huge advances in a branch of Machine Learning called Deep Learning. This describes a family of algorithms based on neural networks. These algorithms are able to learn efficiently from example, and subsequently apply this learning to new data. Here, Zvika Ashani explains how deep learning technology can boost video surveillance systems.

Relationship between deep learning and artificial intelligence

There have been huge advances in a branch of Machine Learning called Deep Learning

With deep learning, you can show a computer many different images and it will "learn" to distinguish the differences. This is the "training" phase. After the neural network learns about the data, it can then use "inference" to interpret new data based on what it has learned. For example, if it has seen enough cats before, the system will know when a new image is a cat. In effect, the system “learns” by looking at lots of data to achieve artificial intelligence (AI). Larry Anderson explores how new computer hardware - the Graphic Processing Unit (GPU) – is making artificial intelligence accessible to the security industry.

Improving surveillance efficiency and accuracy with AI

The Graphic Processing Unit (GPU) – is making artificial intelligence accessible to the security industry

Larry Anderson explains how the latest technologies from Neurala and Motorola will enable the addition of AI to existing products, changing an existing solution from a passive sensor to a device that is “active in its thinking.” The technology is already being added to existing Motorola body-worn-cameras to enable police officers to more efficiently search for objects or persons of interest. In surveillance applications, AI could eliminate the need for humans to do repetitive or boring work, such as look at hours of video footage.

Intelligent security systems overcome smart city surveillance challenges

In surveillance applications, AI could eliminate the need for humans to do repetitive or boring work, such as look at hours of video footage

AI technology is expected to answer the pressing industry questions of how to use Big Data effectively and make a return on the investment in expensive storage, while maintaining (or even lowering) human capital costs. However, until recently, these expectations have been limited by factors such as a limited ability to learn, and high ongoing costs. Zvika Ashani examines how these challenges are being met and overcome, making artificial intelligence the standard in Smart City surveillance deployments.

Combining AI and robotics to enhance security operations

Artificial intelligence is becoming the standard in Smart City surveillance deployments

With the abilities afforded by AI, robots can navigate any designated area autonomously to keep an eye out for suspicious behaviour or alert first responders to those who may need aid. This also means that fewer law enforcement and/or security personnel will have be pulled from surrounding areas. While drones still require a human operator to chart their flight paths, the evolution of artificial intelligence (AI) is increasing the capabilities of these machines to work autonomously, says Steve Reinharz.

Future of artificial intelligence in the security industry

While drones still require a human operator to chart their flight paths, the evolution of artificial intelligence (AI) is increasing the capabilities of these machines to work autonomously

Contributors to SourceSecurity.com have been eager to embrace artificial intelligence and its ability to make video analytics more accurate and effective. Manufacturers predicted that deep learning technology could provide unprecedented insight into human behaviour, allowing video systems to more accurately monitor and predict crime. They also noted how cloud-based systems hold an advantage for deep learning video analytics. All in all, manufacturers are hoping that AI will provide scalable solutions across a range of vertical markets.

Download PDF version

People mentioned in this article

Larry Anderson

Zvika Ashani Agent Vi (Agent Video Intelligence)

Steve Reinharz Robotic Assistance Devices, LLC.

In case you missed it

What are the physical security challenges of smart cities?

The emergence of smart cities provides real-world evidence of the vast capabilities of the Internet of Things (IoT). Urban areas today can deploy a variety of IoT sensors to collect data that is then analysed to provide insights to drive better decision-making and ultimately to make modern cities more livable. Safety and security are an important aspect of smart cities, and the capabilities that drive smarter cities also enable technologies that make them safer. We asked this week’s Expert Panel Roundtable: what are the physical security challenges of smart cities?

Access control systems: Ethernet vs proprietary bus network cabling

When designing a security system for a site, the question of how it should be interconnected is often one of the first you need to answer. Should you choose a system that has its own proprietary bus network, which might require twisted pair cabling, or perhaps one based on an ethernet backbone? Both types of network have their advantages and disadvantages as discussed below. Ethernet connectivity Some security systems are based on a number of modules, and each module is connected to its own ethernet connection. One big advantage of a system like this is that, in many cases, it can be much more convenient, allowing the installer to utilise existing network cabling and other infrastructure, rather than needing to install new cabling. On the other hand, if a security system relies entirely on networking infrastructure controlled by others, typically the IT department, then the stability and reliability of the security system is dependent on that network being available when your system needs it. The stability and reliability of the security system is dependent on that network being available Another potential disadvantage is that certain areas of the premises may not be equipped with a nearby network outlet, and if the network in question is not managed by you, it might be necessary to request the IT department add an outlet for you to use. Proprietary bus connectivity A system with its own proprietary bus network can also have advantages. Perhaps the first and most important difference is that, because the network cabling is installed specifically for the security system, the designer has the luxury of being able to decide exactly where the wiring should be placed and terminated. Another advantage is that the cabling would only be used by the security system, so the installation company can be sure the network will always be available, and there would be very little chance part of it could be accidentally unplugged. Another potential advantage is that some systems are able to run bus cabling of distances well over 1,000 metres, whereas individual ethernet connections are typically limited to 100 metres or less. Another consideration, which applies particularly to intruder and holdup alarm systems, is that communications between elements of the security system should not be prevented by other factors, such as a power failure. Obviously, if a part of such a network is formed by ethernet infrastructure, such as network switches and/or media converters, then that infrastructure needs to be battery backed, and the power supply must be monitored. In some cases, the equipment must be able to withstand a power failure of 24 or even 60 hours. Such long standby times are unusual in IT infrastructure, but are quite common in the case of security systems. The equipment must be able to withstand a power failure of 24 or even 60 hours How this all fits together When selecting a system, it is usually most helpful to have a flexible system that can support a number of different deployment options. This is especially true if the system in question can support a combination of different interconnection types. For example, a single system that can contain a variety of interconnections can then be deployed in a very wide variety of systems where existing infrastructure may be used to aid in the design and deployment: Fibre connections – Many modern sites are pre-cabled with existing fibre connections which can be used to form a dedicated interconnection between system components which can be of the order of kilometres apart. Ethernet connectivity – With the increasing ubiquity of networking within premises, some elements of a security system can be deployed using the existing infrastructure. Repeater - For very large or densely packed systems, a device that can be used as a form of “repeater” can be extremely useful to permit very long interconnect cabling distances. Systems can be formed by utilising a fusion of all of the above connectivity methods Some security systems can be set up to enable multiple discrete access control modules to be deployed, connected to an existing ethernet network, and treated as a single ‘system’ by the management software, while retaining full offline functionality in the event the network becomes unavailable. Further, some systems can be formed by utilising a fusion of all of the above connectivity methods. In practice, of course some applications would suit a deployment that relied solely on ethernet connectivity. Some other applications, especially systems or parts of systems that are part of an intruder and/or holdup alarm system, would better suit a deployment using a dedicated proprietary bus network, and other systems would suit a combination of these communications options. Selecting a system that can be deployed in a variety of ways can be enormously helpful in providing the flexibility projects might demand.

Functionality beyond security: The advent of open platform cameras

The coronavirus (COVID-19) pandemic marks the biggest global disruption since World War II. While the ‘new normal’ after the crisis is still taking shape, consumers are apprehensive about the future. According to a recent survey, 60% of shoppers are afraid of going grocery shopping, with 73% making fewer trips to physical stores. Returning to the workplace is also causing unease, as 66% of employees report feeling uncomfortable about returning to work after COVID-19. Businesses and employers are doing their best to alleviate these fears and create safe environments in and around their buildings. This also comes at tremendous costs for new safety measures and technologies – including updates to sanitation protocols and interior architecture – that protect against COVID-19. Costs in the billions that most businesses will face alone, without support from insurance and amidst larger macroeconomic challenges. Saving costs and increasing security But what if building operators, retail shop owners, and other stakeholders could save costs by leveraging new functionality from their existing security infrastructure? More specifically, expanding the use of current-generation security cameras – equipped with AI-driven image analysis capabilities – beyond the realm of security and into meeting new health regulations. This is exactly where video analytics algorithms come into play. And in the next step, a new evolutionary approach towards open security camera platforms promises new opportunities. Security cameras have evolved from mere image capturing devices into complex data sensors Over the past decade, security cameras have evolved from mere image capturing devices into complex data sensors. They provide valuable data that can be analysed and used in beneficial ways that are becoming the norm. Since 2016, Bosch has offered built-in Video Analytics as standard on all its IP cameras. On one hand, this enables automated detection of security threats more reliably than human operators. And on the other hand, video analytics collect rich metadata to help businesses improve safety, increase efficiency, reduce costs, and create new value beyond security. Expanding camera functionality beyond security Today, we have ‘smart’ security cameras with built-in video analytics to automatically warn operators of intruders, suspicious objects and dangerous behaviors. The rich metadata from several cameras on the same network can also be consolidated by making use of an intelligent software solution. It offers so-called pre-defined widgets to provide business intelligence by measuring area fill levels, counting building occupancy and detecting the formation of crowds. In combination with live video stream data, these insights enable heightened situational awareness to security operators. What’s more, operators are free to set their own parameters – like maximum number of occupants in a space and ‘off limit’ areas – to suit their needs. These user-centric widgets also come in handy in dealing with the coronavirus pandemic. Specific widgets can trigger an alarm, public announcement or trigger a 'traffic light' when the maximum number of people in a space is exceeded. Building operators can also use available intelligence such as foot traffic ‘heat maps’ to identify problem areas that tend to become congested and place hand sanitiser stations at heavily frequented hotspots. At the same time, the option to perform remote maintenance on these systems limits the exposure of technicians in the field during the pandemic. Again, the underlying camera hardware and software already exist. Cameras will be able to ‘learn’ future functionality to curb the spread of the coronavirus Looking ahead, cameras with video analytic and neural network-based analytic capabilities will be able to ‘learn’ future functionality to curb the spread of the coronavirus. For instance, cameras could monitor distances between individuals and trigger voice announcements when social distancing guidelines are violated. Facial recognition software can be trained to monitor personal protective equipment (PPE) compliance and sound alerts for persons entering buildings without masks. The technical requirements are already in place. The task at hand is to deliver these new functionalities to cameras at scale, which is where open camera platforms hold the key. Why open camera operating systems? When it comes to innovating future camera applications that extend beyond security, no hardware manufacturer should go at it alone. Instead, an open platform approach provides the environment for third-party developers to innovate and market new functions. In essence, an open platform principle allows customers and users to change the behavior of devices by adding software afterwards. This software can either be found in an app store or can be self-developed. For a precedent, we can look at the mobile phone industry. This is where software ecosystems like Android and Apple’s iOS have become the norm. They have also become major marketplaces, with the Apple App Store generating $519 billion in billings on 2019, as users use their phones for far more than just making phone calls. In the same way, intelligent cameras will be used far beyond classic video applications in the future. To get there, adopting an open platform principle is essential for a genuine transformation on an industry level. But establishing an open platform principle in the fragmented video security industry demands a cooperative approach. In 2018 Bosch started a fully owned start-up company, Security & Safety Things, and became one of five founding members of OSSA (Open Security & Safety Alliance). With more than 40 members, the Alliance has collectively created the first Technology Stack for “open” video security devices. This includes the OSSA Application Interface Specification and Compliant Device Definition Specification. An open camera platform for innovating future functionality Based on OSSA’s common APIs, collective approach on data security and core system requirements for video security cameras, the first camera manufacturers were able to build video security cameras that adopt an open platform principle. Further fueling innovation, OSSA focused on driving the creation of one centralised marketplace to unite demand and supply in the market. Camera devices that are built in accordance with OSSA’s Technology Stack, so-called “Driven by OSSA” devices, can benefit from this marketplace which consists of three pillars: a development environment, an application store, and a device management portal. Security & Safety Things has advanced OSSA’s open camera platform concept, built this marketplace for the security and safety industry and has developed the open OS that powers the first “Driven by OSSA” devices. Making it quick and simple to customise security solutions by installing and executing multiple apps This year, Bosch, as one of the first camera manufacturers, introduces the new INTEOX generation of open platform cameras. To innovate a future beyond security functionality, INTEOX combines built-in Intelligent Video Analytics from Bosch, an open Operating System (OS), and the ability to securely add software apps as needed. Thanks to the fully open principle, system integrators are free to add apps available in the application store, making it quick and simple to customise security solutions by installing and executing multiple apps on the INTEOX platform. In turn, app developers can now focus on leveraging the intelligence and valuable data collected by analytics-equipped cameras for their own software developments to introduce new exciting possibilities of applying cameras. These possibilities are needed as smart buildings and IoT-connected technology platforms continue to evolve. And they will provide new answers to dealing with COVID-19. The aforementioned detection of face masks and PPE via facial detection algorithms is just one of manifold scenarios in which new apps could provide valuable functionality. Contact tracing is another field where a combination of access control and video analytics with rich metadata can make all the difference. Overall, open camera platforms open a future where new, complex functionality that can save lives, ensure business continuity and open new business opportunities will arrive via something as simple as a software update. And this is just the beginning.