Gaining an Edge: How Embedded Analytics Enable Proactive Security and Business Solutions

surveillance camera
William Brennan
William Brennan

During the past 10 to 15 years, a savvier video surveillance end user has emerged demanding solutions that feature high-performance cameras, versatile installation, flexibility to expand, cost efficiency in storage and advanced analytics to help realize a system’s full potential. Increased public and private security concerns have driven a technology shift and altered business models in the video surveillance marketplace, while end-user demands have forced vendors to conform to their needs.

According to IHS Research, the surveillance market’s migration from analog to IP video systems continued to accelerate in 2018, when 70 percent of all security cameras shipped were network cameras. At the same time, global shipments of HD CCTV cameras, also known as analog HD cameras, fell in 2018. And that trend, which began in 2014 when IP cameras first overtook analog in total sales, is expected to increase.

The rapid growth of IP video surveillance has been driven by end users’ needs to expand the capabilities of traditional CCTV beyond simple monitoring of facilities and assets. While the market demand for smaller and less robust analog systems remains, an increased appetite for next-generation solutions powered by artificial intelligence (AI) and machine learning, along with advanced analytics and cloud computing, has broadened the capabilities of video surveillance beyond just security. Attention to business metrics and an expanded systems team that may include facilities and IT has added new uses like posture and demeanor detection, traffic monitoring and heat mapping, people counting and facial recognition.

The evolution of commodity hardware into a convergence of advanced software and interoperable hardware has allowed systems designers and systems integrators to meet the expectations of end users. This convergence has moved systems solutions from being reactive to proactive, permitting users a seamless integration of technology no matter what level of investment. For example, a relatively small system of 50 cameras or fewer could provide basic video surveillance solutions at a cost-effective price point that would include video management software licensing and servers for 30 days of retention as a technology bundle.

A similar scenario unfolds for users who move into the mid-level systems of 50 to 150 cameras. Now, however, the system may feature more advanced solution-driven devices such as multi-sensor cameras or pan-tilt-zoom cameras that possess higher retention times. Mid-level system users may even request fail-over options for their servers as an upgrade. These servers match the exact specifications that run the main system, but in the event of a power surge or loss of power to the main server, functionality would automatically jump to the fail-over server in a seamless transition.

Advanced Video Technology Raises the Bar

As the digital revolution alters the landscape of video surveillance and security rises to a board-level priority for private and public organizations, security leaders have an opportunity to enhance risk management. Both traditional on-site surveillance and integrated remote security solutions are benefitting from the proliferation of AI-driven proactive solutions that push intelligence to the edge with embedded analytics. This has been a game changer not only for security, but for business as well, altering how video surveillance systems are designed, configured and operated.

According to a 2018 “Future of Physical Security” survey conducted by Microsoft and Accenture, security/risk executives identified “reactive threat management” and “intuition-led decision-making based on subjectivity” as the two leading challenges facing physical security operations today. These challenges – reactive operations and sub-optimal decision making – make it difficult to be proactive. Operating in this manner puts an organization’s people, brand and reputation at risk.

The survey said that, in an ideal environment, security end users would possess the tools to proactively assess and manage their security risk. However, in many cases, the complexity of incoming threats and the sheer volume of information overwhelms security personnel who still rely on manual processes to translate data, which can then inhibit their ability to focus on proactive threat management. The survey added that it is nearly impossible to monitor all security video content without advanced analytics, which is “why it is commonly estimated among security professionals that more than 90 percent of security video footage goes unseen” and video is typically watched only during reactive investigations.

To that end, more than 80 percent of surveyed security participants identified big data and analytics as a top-three investment for the next three to five years. They also said their teams will concentrate on technology investments in cloud computing and storage (58 percent) and advanced identification (56 percent), concluding that a digital video solution “empowers these operators with systems that contextualize data to identify threats before they occur, mitigate risks and better ensure life safety.”

It seems that 2019 became the year that AI technology in video surveillance established itself as mainstream. In a market report released last year by Research and Markets, experts stated that significant improvements are being made in AI video analytics software, and over the next 10 years, it will become a standard feature in most surveillance solutions. The research added that “there is a critical need to make full use of the massive amounts of data being generated by video surveillance cameras and AI-based solutions are the only practical answer. Modern chip architecture with AI software can comb through vast volumes of data and boost security and safety. Granted, there is a lot of development in this field that we are yet to see, but the path towards AI seems quite clear.”

Make no mistake as to what is driving this rapid technology advancement: intelligent hardware and software that allows users to put all of this data to work in a proactive manner. The migration of intelligence to the edge has dramatically increased situational awareness and created an ecosystem of predictive analytics that not only makes for more robust security operations but also a healthier bottom line.

Meeting End User Expectations

End user expectations will continue to push video surveillance to the edge – literally. Video surveillance has traditionally been a tool of the physical security community, who could monitor or recall video footage to identify suspicious activity. However, legacy and closed systems were limited and could only assimilate and translate so much data. The advent of AI technology provides the security user a machine partner that can learn, identify and interpret suspicious activity.

As James McHale, the managing director and founder of Memoori Research, put it, “AI is a tool of the IT community and offers a far more powerful video surveillance solution than traditional approaches. Meaning the security department needs to develop IT skills, utilize the skills of the IT department, or ignore a technology that is changing the physical security landscape.”

Oddly enough, it has been product innovation on the consumer technology side that has helped spur a revolution in the security industry as the integration of advanced analytics and edge computing in everything from smartphones to smart refrigerators shapes the burgeoning Internet of Things (IoT). Video data processed at the edge – that is, in the camera itself – lowers bandwidth consumption and reduces communication latency, allowing end users to create more proactive security strategies and take appropriate action in critical situations.

End users appreciate the lead time provided by the rapid transmission of data offered by edge computing, especially considering that it takes 150 to 200 milliseconds for data to travel from endpoints to the cloud and just 10 milliseconds from endpoints to the edge. This speed allows for more effective detection and response.

Lower bandwidth consumption is another major benefit of edge computing as all the processing is done at the source of the data. While bandwidth issues have improved for many organizations, applying advanced analytics on video that runs through a central server and is, thus, heavily compressed, reduces the accuracy of the analytics. Moving the processing of the original data to the edge is an obvious advantage when it comes to response time and image resolution.

Camera latency is a major concern for end users, especially in critical infrastructure and gaming environments. Edge computing reduces latency since video does not have to be sent to the back end for processing and analysis. This has significant advantages in applications like facial recognition, where detection of significant situations at the edge triggers an alarm response from the device itself, skipping the round-trip transmission to the back-end servers. There are also aspects of cloud computing, such as “fogging,” that can extend cloud computing to the edge of a network while also placing data, compute, storage and applications in the most logical and efficient location between the cloud and the origin of the data itself.

What Drives Technology?

Technology continues to develop and come to market at an incredible pace driven by new intelligent hardware and software with AI that deliver improved outcomes. For example, new IP camera systems with embedded analytics and sensors are capable of detecting far more than a conventional camera (or human eye) can see. These intelligent edge devices dramatically increase situational awareness for conventional security applications through predictive analysis capabilities, while providing reams of data for business intelligence applications. Meanwhile, advanced algorithms and analytics like facial recognition and various forms of object detection further capitalize on the powerful imaging and detection capabilities of edge devices.

Physical security is only as good as the intelligence the technology offers. End user demands, now more than ever, are helping to chart the course for the technology advances of the future. For years, the security industry strived to achieve networkability, and now that it has arrived, it is dramatically changing the dynamics of the industry. In this age of IoT, users are looking for interoperability and open architecture systems. Users want to converge systems and integrate everything, which has inspired many companies to think outside the box and develop more easily integrated and operated system solutions. Where cameras were once deployed simply to monitor activity and document events, they are now being used to detect abnormalities, count people, monitor motion direction and speed and so on.

Users want proactive performance that captures and analyzes more data for applications that can help them improve merchandising on retail floors, increase manufacturing throughput, detect traffic violations, get alerted to unauthorized site visitors and more. As users demand more performance, versatility and intelligence from their systems, the opportunity to fulfill their requests creates new growth opportunities for solutions providers.

William Brennan is vice president of the security division for Panasonic i-PRO Sensing Solutions.