According to the World Health Organization, by 2050 seven out of every ten people will live in a city – an urban population of 6.4 billion.[1] This explosion of population density means that public service organisations will face complex safety and security challenges. Knowing what is happening and where helps, and most city dwellers have accepted cameras’ watchful eye as a fixture of their urban habitat. But the sheer volume of data being captured means that human analysis is simply not fast or thorough enough to make best use of it; indeed, some estimate that up to 98% of CCTV footage remains unseen by anyone.[2]
Two core challenges are affecting the effectiveness of CCTV as a public safety solution: First, public safety challenges are creating ‘the perfect storm’ – a combination of budget cuts, evolving threats, decreased public safety officer headcount and increased citizen expectations, highlight the need to develop and maintain a cost-effective response. Used intelligently, CCTV has the power to address these pressures.
Poor compatibility and integration
A second challenge lies around the fact that CCTV alone does not deter crime. While there is no shortage of cameras, the existing infrastructure must initiate a response to best utilise camera output. The United Kingdom is one of the most monitored countries in the world with an estimated four million cameras nation-wide – but as Detective Chief Inspector Mick Neville of London’s Metropolitan Police says: “The police simply haven’t got the time to watch hours and hours of footage for a theft.”[3] Despite more than 13,000 cameras in Paris, they were ineffective in the capture of the perpetrators of the Charlie Hebdo massacre[4] because they were set up for traffic enforcement and were incompatible with systems that may be used for wider public safety purposes (such as facial recognition or vehicle identification).
The digital age, prompting a vast increase in the volume of video, enhanced connectivity and speed of data exchange, extends the ability to improve the use of CCTV as a public safety solution. However, poor com-patibility between CCTV networks across agencies and non-existent integration of public safety systems and technologies, such as drones and body-worn cameras, is adding to the complexity of managing public safety concerns. With a high number of incidents across vast populations, it is essential to process information swiftly and generate the right intelligence in a timely way; for example, it still took five days to identify the Boston bombers’ rucksack, despite the clear video footage of the incident.[5]
How video analytics can make the most of camera investments
Video analytics can support and enhance the traditional use of CCTV as both a proactive and reactive public safety tool. Video analytics can quickly analyse, alert and report on CCTV footage – even monitor feedback from multiple cameras constantly – without the errors of manual intervention.
Face recognition technology
Video analytics can deliver improved results as part of post-incident analysis. Immense improvements have been made in the accuracy of both video and biometrics analysis in the last five years. For instance, face recognition technology is more accurate than human analysis as vehicles and people can be counted with more than 97% accuracy. Also, age and gender can be automatically assessed with more than 90% precision rates.[6] As a result, video analytics is becoming a far more reliable and trusted technology which can help to free up public safety staff for more valuable tasks – and provide public safety solutions that better serve citizens’ needs.
City centres are particularly vulnerable and often a target for criminals seeking to take advantage of the cover of population density. Here, digital intelligence can help once more. Accenture’s ‘face in the crowd’ recognition solution has been used successfully in a large European shopping centre to scan up to 15,000 visitors per day, and generate notifications for known individuals or persons of interest with less than 0.5% false alerts.
Safe city systems
Although post-incident analysis may be one of the most high-profile applications of CCTV, the majority of the video analytics applications are for counting purposes only. Operations such as transportation and major event management can be optimised and richer operational insights can inspire new ways to serve the public good. Using many different digital technologies and advanced data analytics, like ‘safe city systems’ are capable of detecting and monitoring a variety of situations, such as urban traffic, disturbances to public order, cleanliness and flooding issues.
Enhanced data privacy
While video analytics can make automatic identification far more reliable, automating video capture can also enhance data privacy. Large-scale data collection pro-tects citizens’ anonymity by assessing patterns, not people. Robust privacy laws mean that anonymised data is the only practical answer for most video ana-lytics applications. Once the vagaries of personal bias have been removed through the automated collection and analysis of data, public safety agencies can better predict an appropriate response and determine how best to deploy resources.
Three cities with a common goal
Whether fighting crime, maintaining public order, or providing emergency services, public service organi-sations need to gather intelligence quickly. By gaining insights from video-enabled data-rich sources that already lie at the heart of the city and improving the collaboration between city stakeholders (which include the police, the municipality, emergency services and transportation authority), public service organisations can not only predict areas of need, but also protect citizens through swift incident prevention, detection and response for greater overall public safety. Singapore, Mons in Belgium and Lille in France have all successfully piloted video analytics. The following case studies illustrate how they improved their situational awareness and made the most of limited resources and public safety assets.
Singapore: Safe city test bed
Through its Safe City programme, the Singapore Government seeks to harness the latest in advanced analytics to complement its public safety solutions, increase security, deliver services more effectively and make the most efficient use of its resources. Accenture collaborated with six Singapore government agencies from the areas of law enforcement, transportation and the environment to pilot its Safe City Solution, supported by the Video Analytics Service Platform.
The solution integrated existing and new video sources into a central analytics platform and ran live tests during two major public events. Through the use of advanced video analytics and user-friendly applications, the test bed successfully helped the government to predict crowd behaviour, coordinate resources, respond to incidents and facilitate collabo-ration among the various agencies. The government can now produce meaningful insights in real time, enabling it to respond more quickly to situations.
Belgium: Mons 2015, public safety solutions pilot
The Mons Police Force, in cooperation with the National Police Force of France, decided to deploy advanced police technology analytic capabilities into video protection systems used in the city. By doing so, the public safety solution intended to increase situational awareness, streamline operations and enhance the response times of the police to public safety incidents throughout the city’s 2015 tenure as European Capital of Culture. By adding an automated video analytics system to the traditional video protection network, the police were able to detect and evaluate more incidents and anomalies, providing a better service to citizens and improving security at the event. In particular, the solution helped to avoid congestion and overcrowding in the main streets of the city and meant the Mons police force could make more informed decisions about access to high traffic areas to improve public safety.
France: A public service agency uses analytics
Affiliated with the French Ministry of the Interior, the Service des Technologies et des Systèmes d’Information de la Sécurité Intérieure (ST(SI)2) is responsible for protecting the public at densely populated events. One such event, the Grande Braderie de Lille, sees the town’s population swell from 230,000 inhabitants to more than two and a half million people over a weekend every September. ST(SI)2 decided to pilot a video analytics service platform at the annual Lille event to increase information accuracy, improve decision making and enhance the performance of the French public safety officials’ operational teams. By adding analytics to standard CCTV footage, police officers could detect and assess far more incidents, provide a better service to citizens (they helped 502 people versus 450 in 2013) and improve safety throughout the event (three adults and 13 children who had been lost were returned to their families). What is more, the police solution connected to existing external public safety systems to create greater intelligence.
Four steps to effective video analytics
The three cities that successfully piloted video analytics had a comprehensive approach. This can be summarised in four best practices and action points for pairing CCTV with automated analysis: redefine your strategies, embrace new partnerships, flex technology and drive a dynamic approach.
Develop a strategy
Using video analytics and making it part of how public safety organisations operate already has the support of citizens. Eight out of 10 citizens surveyed declared they were in favour of police using digital technologies – 83% citing they are comfortable with CCTV.[7] Video analytics can benefit police officers in the field – improving situational awareness, providing a tactical information feed and conducting real-time assess-ments. There is evidence, too, that body-worn cameras can have a positive effect on behaviours. When officers and civilians are aware their actions are being recorded, both are likely to act in a more civil manner – the Mesa Police Department found 77% of police officers believe body-worn cameras cause them to behave more professionally. Finally, video analytics and cameras are also a useful training tool, helping to prepare officers and improve the quality of service by providing a comprehensive view of incidents and interactions.[8]
Action: Shape a strategy for the next five years that builds video analytics into your organisation, taking account of specific operational needs and demands. The strategy should also assess the implications for and practicalities around the storage and management of video data.
Partner and interact in new ways
Public safety organisations can draw on the resources and capabilities of the private sector, either to gain greater intelligence or work collaboratively on a tech-nology solution. Accenture collaborated with six govern-ment agencies to pilot the Singapore government’s Safe City Solution. The government can now integrate public safety video information from a range of agencies and produce meaningful insights in real time.[9]
Action: Find new ways to interact with the private sector to gain access to their networks, take advantage of implementation support for video analytics, and benefit from private sector resources in the editing and packaging of CCTV footage for public safety agencies.
Flex your technology
Video analytics algorithms evolve rapidly in terms of capability and accuracy. In one year, an algorithm being used today may no longer be the most effective; as a result, the solution must be flexible to cover diverse situations and integrate the latest advancements. Compared with traditional storage options, cloud solutions offer an opportunity to store a vast amount of data collected by CCTV in a cost-efficient manner. Used effectively, the results from a combination of CCTV and video analytics can provide additional alerts at a lower cost.
Action: Maintain a flexible platform by adopting an integrated and component-based solution, including considering flexible options such as cloud-based storage, that is able to adapt and take advantage of the best combination of technologies and reduce costs.
Drive a dynamic, multi layered approach
Video analytics is not a single solution but a dynamic and changeable combination of inputs. By analysing multiple sources of footage from fixed point CCTV, private sector videos, mobile CCTV and body-worn cameras – not to mention the growing video evidence filmed by the public on their own mobile phones – public safety agencies are better placed to serve the safety and security demands of their local communities. In this way, a machine-based system can become the ‘eyes’ and supplement the resources of public safety workforces. By adding an automated video analytics system to the traditional video protection network, the Mons police in Belgium were able to detect and evaluate more incidents and anomalies during the celebration of the cultural capital of Europe opening ceremony.[10]
Action: Combine diverse video footage from multiple sources using video analytics to add new breadth and depth to public safety information that inspires real-time decision making.
Where next?
Arguably the biggest driver of productivity and growth for many organisations in the next decade, the Internet of Things – where existing objects are connected to the Internet to collect and receive data – will affect the private and public sector alike.
Business models, workforces and how data is shared are all being reinvented by this latest wave of digital innovation. For public safety agencies, this could mean that when a police officer enters a dangerous zone, his presence is automatically detected, activating the public or private CCTV cameras to offer support, cover or intelligence about the location.
The evolving sophistication of video analytics means that its accuracy and relevance are also growing. For example, video analytics could be used to detect violence or monitor prisoner congregations by sending alerts to pre-empt fights or assaults, leading to safer prisons and jails.
Conclusion
To understand how safe a city is, it is essential to monitor it daily and feed that vital intelligence back into creating further improvements and maintaining consistent progress. Under pressure to do more with less, municipalities and local governments would do well to embrace digital technologies and advanced analytics to better manage day-to-day city operations and enhance emergency response. By redefining their strategies, embracing new partnerships, flexing tech-nology and driving a dynamic approach, public safety agencies can take advantage of the new, improved video analytics to deliver public service for the future.
References
1 World Health Organization. http://www.who.int/kobe_centre/publications/hiddencities_media/p1_who_un_habitat_hidden_cities.pdf
2 Video Analytics and Privacy: A Way Forward. Accenture, 2014.
3 Mick Neville at ST13 Newcastle. Professional Security magazine, 2013. http://www.professionalsecurity.co.uk/news/interviews/mick-neville-at-st13-newcastle/.
4 Tucker, P. (2015). Here’s Why Security Cameras Were No Help In Capturing Paris Terrorists. Defense One. http://www.defenseone.com/threats/2015/01/heres-why-security-cameras-were-no-help-capturing-paris-terrorists/102438/.
5 Balderson, K. (2013). Color of Alleged Boston Bomber’s Rucksack Contradicted? WideShut. http://wideshut.co.uk/color-of-alleged-boston-bombers-rucksack-contradicted/.
6 National Institute of Standards & Technology. http://www.nist.gov/index.html.
7 How can digital police solutions better serve citizens’ expectations? Accenture, 2014.
https://www.accenture.com/t20150904T051924__w__/us-en/_acnmedia/Accenture/Conversion-Assets/DotCom/Documents/Global/PDF/Dualpub3/Accenture-How-Can-Digital-Solutions-Better-Serve-Citizens-Expectations.pdf.
8 Body Worn Cameras have arrived; now comes the hard part. Accenture, 2015.
https://www.accenture.com/t20150523T024229__w__/us-en/_acnmedia/Accenture/Conversion-Assets/DotCom/Documents/Global/PDF/Dualpub_14/Accenture-Bodyworn-Cameras-Weis.pdf.
9 Singapore Government: Safe City Test Bed. Accenture, 2014. https://www.accenture.com/t20150707T195238__w__/us-en/_acnmedia/Accenture/Conversion-Assets/DotCom/Documents/Global/PDF/Industries_13/Accenture-Singapore-Government-Safe-City-Test-Bed.pdf.
10 Mons 2015, European Capital of Culture: Public safety solutions pilot. Accenture, 2015. https://www.accenture.com/
Joost van der Burg is Senior Manager at Accenture and works within the company’s Defense & Public Safety business. Within this position, Joost is actively pursuing opportunities with regard to Video Analytics. Joost has over 15 years of experience within this specific area of expertise (Police, Prison Services, Ministry of Security & Justice and the Ministry of Defense).
