Trends, likely developments, upcoming projects, opinions, how the industry sees itself, legislation, and the influence of industry bodies are the basis for our survey
Jeremy Malies, content editor at MEB Media, identifies three key factors that will likely shape the camera surveillance market in the next 12 months. His areas of focus are artificial intelligence (predictable but inescapable as a topic and will naturally predominate), platform aggregation, and green credentials which now often form part of corporate citizenship.
The article also looks at three dynamic and significant vertical sectors for camera manufacturers and installers. Jeremy identifies these as transport hubs (particularly airports), the built environment (including heritage architecture) and custodial premises.
In Security assesses offerings from major players in the camera surveillance sector, and canvasses opinions from executives at these companies.
Artificial Intelligence
So how does artificial intelligence show itself in camera surveillance, how is it being harnessed and – the biggest question that analysts ask across technology sectors – what is artificial intelligence learning for itself and might it become truly conscious?
In terms of the big deployment of AI in camera surveillance, the answer to what the technology is learning is everything. And that is, of course, at the centre of how analytics algorithms work. The software learns to discount irrelevant movement in a scene and instead alerts on atypical activity which might be an intruder trying to breach a fence.
Any technology writer would be foolish to ignore artificial intelligence in a survey of this kind. Having worked at a science park in Cambridge in the late Nineties, I can remember that the software house that came up with the first algorithms to automate decisions on what constituted a threat described them as “intelligent scene analysis” which to me is a better phrase than “video analytics”.
Responsible camera developers do not reach immediately for the phrase “artificial intelligence”. They are more circumspect. But manufacturers will use the phrase “deep learning” and this features in the literature of MOBOTIX who have been associated with high-end video systems since 1999.
Edge analytics and process optimization
MOBOTIX (part of the CERTINA Group) is synonymous with the decentralized edge approach to camera surveillance by which much functionality, including the ability to discriminate between (and alert on) different objects and behavioural patterns, sits in the camera itself rather than at the core.
The MOBOTIX analytics excels in security environments but has moved beyond these to analyze staff and visitor movement for general logistical purposes. Shop owners or exhibition organizers can run reports that tell them which part of a store or tradeshow floorspace is generating most interest, which part of an airport departure hall is preferred by travellers etc.
Capturing and evaluating staff and customer movement patterns (heatmaps are an effective tool) is not true artificial intelligence as a rigorous interpretation of the phrase would demand, but it shows automated decisions being made or suggested as people and objects are tracked and footfall counted. Tracking area occupancy and movement styles also means that camera surveillance can improve employee safety.
Alerts on unexpected movement patterns within defined areas are now allowing MOBOTIX to offer fall detection by which artificial intelligence reports on what might be a trip by a senior, a post-operative convalescing patient, or anybody whose mobility and balance are impaired. Detecting and mitigating the effect of falls through alerts that bring prompt intervention shows MOBOTIX contributing to the healthcare sector.
Finding patterns
“Finding patterns” is a phrase in the literature of Hikvision on intelligence in cameras that struck a chord with me. Again, the Chinese manufacturer stresses the importance of “deep learning”, and they are not shy of using the term “artificial intelligence”. The cameras use machine learning to gather insight on what they see at a scene and then discount anything that the intelligence understands to be ambient behaviour in order to alert on truly atypical patterns of movement and object types that may constitute a threat.
The cameras will also find relationships within data to give reports to managers in non-security logistical roles. Again, analysis is performed at the edge with objects being categorized and assessed. Queue management and perimeter protection are particular strengths of the edge algorithms.
Hikvision’s artificial intelligence has the versatility to work with and complement smoke detection systems. A recent alliance has seen Hikvision collaborate with Belgian company iThermAI. Together, the partners are generating alarms on live feeds from security cameras before there is any contact with heat or smoke. iThermAI algorithms have been embedded into Hikvision 4K and bi-spectrum cameras.
The partnership underlines what we often dismiss as a platitude from the organizers of trade fairs; the security and fire protection sectors are indeed becoming more and more aligned. It also shows how patterns within threat scenarios that usually require human intelligence (always liable to distraction or fatigue) can be automated with rapid response times that are beyond human capability.
Hanwha Vision
In far-reaching analysis of what AI may hold out for the security sector, Hanwha Vision has noted how video management systems (Cloud-based is the most effective platform) are bringing analytics beyond algorithms to interfaces where the user can ask questions of the video storage platform using intuitive natural language.
A security guard could simply ask the VMS to interrogate footage over the last 24 hours for even a commonplace scenario such as a visitor in black clothing interacting with a receptionist at a particular time. Of course this is not the same thing as consciousness, but AI cameras will soon have a rudimentary understanding of situations such that they can assess what general movement of people on site may mean and so generate real-time alerts.
Hanwha Vision are putting forward the idea (credibly) that analytics will soon move beyond the current scenarios of being able to detect running, loitering or climbing and be able to evaluate such movement as a series or development of events so showing the underlying cause and significance.
Andy Ryu, Chief Product Officer (Vice President) at Hanwha Vision, said, “This technology [integration of generative AI and re-identification of a single person or object] has far-reaching applications beyond criminal investigations and missing person searches. It can revolutionize traffic management and disaster response and even provide valuable insights into visitor behaviour for optimizing operational efficiency and gathering marketing intelligence.”
As with the collaboration described above between Hikvision and iThermAI, Hanwha Vision sees a new era of openness and joint initiatives between companies in an expanded AI ecosystem. A willingness to work together may become an almost Darwinian measure of adaptability.
The company even likens a pick-and-mix selective attitude in deploying AI solutions to the way we choose Apps for our smartphones. Recognition of the need to make a whole ecosystem of artificial intelligence Apps available to users has been shown in Hanwha Vision’s recent release of a four-channel AI PTRZ multi-directional camera.
City of Bonn case study
Hanwha Vision has a transport installation in western Germany. The project underlines much of the analysis above and shows the company’s products protecting residents and tourists across bus routes, train routes and a tram network for transport operator Stadtwerke Bonn Verkehrs-GmbH. A combination of more than 600 4K IR outdoor vandal dome and pan-tilt-zoom cameras has been installed by multi-technical services integration partner SPIE.
The locations include tunnels, parking facilities, entrances/exits, walkways, escalators and lifts. Benefiting from AI alerts, the cameras report to a central control room so empowering staff to distinguish between nuisance alarms and genuine events/emergencies. Hanwha Vision’s deep-learning algorithms are allowing the transport network staff to receive filtered information on object detection and classification including people, vehicles, facial images and number plates. Shadows and wildlife are discounted with high levels of accuracy. The embedded camera intelligence is also used for people-counting that can warn against and prevent bottlenecks (so stopping overcrowding risks) as well as creating alerts on trespass and genuine incidents of trip & fall.
So what is the artificial intelligence landscape in 2025?
More precise motion detection in the next year appears to be a given. This is all of a piece with agentic artificial intelligence, this being the class of AI whereby systems act autonomously and do not look for human supervision. Vendors boast of ability to process huge amounts of data (it might be a high-resolution video stream) with in situ adaptation to changes in the environment. The intelligence even approaches the level of strategic planning and dynamic problem-solving.
Object recognition appears to be an obvious task to be treated in this manner. It will become more sophisticated, and algorithms will be quicker to alert operators to anomalies. The more that routine monitoring can be automated, the more focus can be given to high-priority tasks by site managers. This frees up resources for strategic decision-making according to the PR blurb of many manufacturers. In other words, artificial intelligence is being used as an augmentation tool that does the by-rote work. Will this mean job losses? I can’t find redundancy announcements within our physical security sector, but earlier this year, CrowdStrike – a cyber security company based in Austin, Texas – laid off 500 people who represented five percent of the workforce. A spokesperson said that this was, “A market and technology inflection point, with AI reshaping every industry.”
Platform aggregation
Walking the floor of The Security Event in Birmingham recently and talking to In Security associates and partners, I learned that end-users are insistent now that they should receive information from the same platform however disparate the information might be in nature.
An aggregated platform and unified management framework are their best opportunity to blend fragmented data and find patterns in it which mean that they make the best possible decisions. Cameras cannot be standalone, access control readers cannot be isolated, and intrusion detection alarms cannot be piecemeal. Solutions need to be end-to-end in order to optimize situational awareness. Vendors, consultants and end-users now liken the platforms to a single pane of glass or an overall intelligent ecosystem.
Camera surveillance now needs to be seen as a proactive, intelligence-driven resource rather than reactive protection measure. The huge range of city-wide sensor devices that form the Internet of Things (this belongs to another article) means that CCTV can be the backbone security asset for an entire city.
A new generation of city planners no longer recognise a traditional model where security installers simply devise and implement systems. Young well-educated civil servants (the type who are not likely to become redundant through artificial intelligence) have grown up with Security-as-a-Service. They look for continuous monitoring with software patches and are comfortable with subscription-based models. They expect physical security to be accessible and scalable, and capable of making a case for itself through cost-effectiveness and long-term cost of ownership arguments. The vendors who prosper will be the ones who can share or at least empathize with this mindset.
To say that physical and cyber security are converging now would be a naïve truism. It is more relevant to say that the two types of security as well as operational technology are part of a unified ecosystem.
Axis Communications and joined-up thinking
This kind of collaborative approach has been illustrated, embodied even, by the Swedish manufacturer Axis Communications with its Axis Cloud Connect. The company’s developers are advancing Cloud-connected solutions with data aggregation of all kinds and transformation of VMS integration.
Scalability, flexibility, innovative video analytics scenarios and robust cybersecurity are all part of the initiative that is seeing collaboration with partners including Eagle Eye Networks, SecuriThings and Wesco. Axis Cloud Connect is giving integrators a new perspective on device lifecycle management and preventative maintenance.
In April 2024 during ISC West, Fredrik Nilsson, VP, Americas, Axis Communications, said, “Over the past year Axis Cloud Connect has helped transform the way security, business intelligence and operational solutions are deployed, utilized, and managed. The open platform’s flexibility, scalability, and advanced cybersecurity features have enabled our partners to innovate and deliver cutting-edge solutions that meet the dynamic needs of the market. The progress made by our partners is a testament to the power of openness and collaboration, and the immense potential of Axis Cloud Connect to shape the future of security and IoT.”
Corporate citizenship and green credentials
If there has been a quantum leap in terms of environmental practice across the security industry it is that companies no longer have to be environmentally aware simply in order to be good corporate citizens or to meet tender requirements from clients who are looking at their own environmental record. Ethical behaviour in terms of carbon footprint often makes good business sense as well.
Minimizing electrical consumption, reusing packaging, and opting for sea rather than air freight can all improve the bottom line. And the perspective has widened beyond obvious factors such as optimum amounts of recycling and avoiding landfill to issues such as reducing water usage and having meaningful numbers of solar panels at manufacturing plants.
Manufacturers who walk the walk
Of the major players surveyed and assessed for this article, Hangzhou-based Chinese manufacturer Dahua Technology can be adjudged to be not only walking the walk in terms of minimizing their environmental impact but also introducing technology that fosters vulnerable wildlife. Dahua held meetings earlier this year with the World Wide Fund for Nature (WWF) to explore collaboration.
Dahua has implemented its Artificial Intelligence of Things (AIoT) in Antarctica and across a rain forest in Hainan which is an island province in the extreme south of China. The province features a wildlife park that is a habitat for giant pandas. Dahua solutions have allowed remote monitoring of pandas and gibbons.
Bob Chen, Senior Vice President of Dahua Technology and President of Global Business, said, “Dahua is committed to working with organisations like WWF, combining our strengths in technology to contribute to ecological conservation and promotion of sustainable practices.”
Dahua’s conservation activity is broad; the company combats illegal logging and uses thermal video monitoring to detect illegal fishing. It even incentivises installers to support conservation projects including preservation of wetland ecosystems.
At the fossil fuel level, the Dahua Smart (IoT) Industrial Park on the Qiantang River is using a smart electricity system that saves up to 1,389,800 kWh of electricity a year. Photovoltaic power stations are being used in order to generate clean energy and so reduce carbon emissions.
Paxton Access
Another company whose green credentials are impeccable (to the point of being carbon-neutral) is Brighton-based manufacturer and developer Paxton Access. I know since I’ve visited their headquarters in order to attend two press days. So, what place do Paxton whose focus is of course access control have in this article?
The company’s heritage architecture case studies include Royal Connaught Park in Hertfordshire which, among many claims to fame, has been a shoot location for the iconic and still re-run constantly today series from the Sixties, The Avengers. It starred Patrick Macnee, Honor Blackman, and Diana Rigg and has a cult following among people of all ages.
While not a major part of the installation (the focus of the project is access control) Paxton10 cameras were installed at the site by Harlow-based integrator Gamma Systems. The camera series includes bullet and dome models, and they are known for requiring zero configuration. Royal Connaught Park was once an Edwardian school that has become a gated residential development of luxury apartments in recent years. Being of a certain age, I gravitated to The Avengers, but it should also be noted that the site has been the real-life shooting location of Hogwarts Great Hall for three Harry Potter films.
Verticals with best opportunities for the industry
With our own industry making real strides in its carbon footprint, what progress is being made in the aviation industry? I ask because while reporting from Intersec in Dubai earlier this year, I could not help but glean information about the extraordinary scale of airport projects in Saudi Arabia – not one of the Emirates of course but subject to the same region-wide macroeconomic forces.
The latest buzzword in aviation is “coopetition”, this being collaboration between rivals for a greater good. The collective goal in this case is greener aircraft both in terms of emissions and consumption of the planet’s finite resources during construction. I’m thinking specifically of CleanAviation, an eleven-strong initiative between manufacturers including Airbus and the European Commission.
King Salman International Airport in Riyadh will be created around the existing King Khalid International Airport and will be the work of architects Foster + Partners (synonymous with Norman Foster) who are based in Battersea, London. There will be no fewer than six passenger terminals and six runways across an area of 22 square miles and with an envisaged annual passenger footfall of 100 million by 2030. Prominent consultants, integrators and manufacturers will no doubt secure the site as the airport takes shape.
KEENFINITY Group airport case study
Meanwhile, at 16 million passengers a year, Changchun Longjia International Airport in north-eastern China is a significant case study for KEENFINITY Group. The end-user required 24/7 recording from over 1,000 surveillance cameras and archiving of that footage for 90 days. The environmental conditions (notably temperature fluctuations and the inevitable presence of fumes and contaminant particles on the tarmac) were a challenge as was the client’s insistence on extreme energy efficiency.
The airport is using 1,500 units of Dinion, Autodome and Flexidome cameras from KEENFINITY and central camera management via BVMS. A small data footprint is being maintained with H.265 compression. An obvious cost-saving is a reduction in the number of storage units that results from the compression technology with positive knock-ons in terms of energy consumption and costs of cooling server rooms. All these benefits are being achieved without any compromise on image quality. Edge algorithms in the cameras are supporting airport security staff with automated alerts.
Heritage built environment
How does the need to keep tourists at heritage landmarks safe using video surveillance coexist with natural scepticism from architecture purists who do not wish to see the outline of period buildings spoilt by bulky surveillance cameras? There have been a few ill-considered CCTV assemblies at famous locations for sure, but the increasingly sleek outlines and compact dimensions of modern cameras are improving things.
Cathedral security
A project where this balance has been achieved comes from the portfolio of IDIS who are a global designer, developer and manufacturer of surveillance products. The company’s case studies stretch the idea of heritage right back to the eleventh century and include an installation at Canterbury Cathedral.
Video surveillance in places of worship can be vital to the safety of all parties, but few sites require more tact and discretion in terms of camera deployment and privacy zones. Using IDIS products, integrators Hall & Kay Engineering balanced a need to secure the site with respect for its sacred nature.
The installation involved IDIS high-definition cameras, recorders, VMS and peripherals. Areas covered included entrance to a welcome centre, retail areas and exhibitions spaces. The IDIS full HD IR dome cameras are vandal-resistant models with true wide dynamic range (WDR) and day and night functionality based on infrared cut filter removal that allows switching between colour and black & white modes.
The constables at the cathedral (a historic term for wardens using a mix of traditional and modern security practice) analyze display from 16-channel NVRs provided by IDIS. There is optimum connectivity of the cameras and mutual two-factor authentication as a cyber security measure. There are also role-specific permissions and user rights for the constables according to seniority within the license-free IDIS Center video management system.
A different kind of tattoo
Like Canterbury Cathedral, the present structure of Edinburgh Castle also dates from the eleventh century. It is part of the security remit of the City of Edinburgh Council which stretches from the city’s half a million citizens to the influx of four million tourists every August for cultural events including the Edinburgh Tattoo held on the esplanade of the castle.
Genetec have an Edinburgh case study as part of the Digital and Smart City strategy of the City of Edinburgh Council. The Council’s innovative strategy for security technology prompted it to approach Genetec to improve public services in terms of safeguarding all who work, live or simply spend time in the city. Liaising with channel partner, North, the Council began to create an intelligent centralized security platform.
This has been in part based on the Genetec Security Center which is an open unified physical platform that allows everything from video surveillance, analytics and IoT sensors to be managed from a single interface. Security Center incorporates automatic license plate recognition (ALPR) and analytics algorithms on many scenarios and combinations of scenarios.
One of the benefits of the Genetec deployment in Edinburgh has been higher quality camera footage which has helped get offenders successfully identified, charged and convicted. Improved footage submitted as evidence has seen an increase in early guilty pleas.
Prisons and other custodial sites
You have to look to the US for notable case studies of video surveillance in prisons. In the UK, use of CCTV tends to be in police custody suites, remand centres and secure mental hospitals. There are privacy issues over in-cell usage of video recording, but no such issues if CCTV is deployed to complement perimeter protection, with thermal cameras being an obvious choice at night. Conscientious system designers make sure that video at prison perimeters can alarm on would-be outside escape assisters entering a sterile zone.
Use of video surveillance with an analytics algorithm to alert on the presence of a drone over the prison grounds (there are frequent attempts to smuggle contraband) would seem an intelligent application in every sense.
Inmates and guards are now aware that (unless a privacy zone is deliberately created) there are few blind spots in video surveillance of a prison. This is valuable in terms of ensuring appropriate behaviour by guards towards prisoners as well as mitigating escape potential and the risk of prisoner violence towards guards or fellow inmates. Unless there is an acute risk of self-harm, prison authorities worldwide show reticence about using video surveillance in cells.
There is a consensus among consultants and prison boards internationally that video surveillance is at its best in creating an alarm on non-violent planned misconduct such as escape rather than inmate-on-guard violence or exercise yard fights. The one environment in which (possibly intrusive) video surveillance is tolerated in order to factor out the risk of suicide appears to be young offender units.
It is a simple matter for analytics to alert staff if an inmate is possibly in distress, with this manifesting as being in a prone position in an unexpected part of the cell. Truly effective use of automated alarms on prisoner behaviour via analytics can help protect prisoner rights by freeing up staff to focus on inmate welfare.
Conclusion
It might be a naïve stance, but my survey here suggests that the time-saving from intelligent analytics and the automation of tasks performed by human operatives is not about to result in large scale redundancies among blue collar staff. Our sector is not the same as cyber security and the company in Texas mentioned earlier. It is more likely that time-saving will result in improvements in visitor experiences and safety levels.
Fewer security guards looking at possible breaches of a perimeter (the kind of scenario that can be automated) means more security guards using their human intelligence to interact with visitors and perhaps frustrate theft, vandalism or even hostile reconnaissance or simply help people who through age, infirmity or technophobia are confused by major retail sites, transport hubs and any campus-style environment.
However tempting it might have been to give artificial intelligence a try-out when writing an article along these lines, this piece truly is one correspondent’s attempt to make sense of current developments in video surveillance!
If MEB Media knows anything about articles written by AI, it is that they are always balanced (excessively) to the detriment of any true opinion being expressed with vigour. So, if the opinions here are thought to be in anyway over-vigorous or ill-considered, our letters column is always available to you!