VCA evolves with latest trends

VCA evolves with latest trends
Video content analytics (VCA) has been touted as the next big thing in video surveillance for over a decade. And while VCA has now undoubtedly become a major part of video surveillance, for a long time it was met with disappointment in the form of inaccuracy and false alarms. Despite this, particularly over the last few years, accuracy has dramatically improved and VCA has become an integral part of a comprehensive video surveillance system, with many cameras nowadays boasting analytics on the edge. In fact, research shows that the reliance on VCA technology is growing along with its adoption. As a result, the video analytics market size is forecast to grow from US$1.7 billion in 2016 to $4.2 billion by 2021, at a CAGR of 20.2 percent from 2016 to 2021, according to a June 2016 report by MarketsandMarkets.

With the market expected to double over the next five years, current user demands for crowd management beyond security applications and collaboration with IoT are helping to drive growth.

Trends in VCA

Trends are constantly coming and going; however, their influence on technology is lasting. Since the introduction of VCA to security, there have been many trends whose technologies are now integrated into what we know as VCA today. Staying on top of new technologies is crucial for VMS developers, as end-user demands follow the latest technological trends.

When asked what is currently pushing VCA development, developers pointed to mobile platforms and crowd management, to name a few.

“Intelligent mobile platforms, such as drones and driverless cars, are a major trend influencing VCA development,” said Eric Olson, VP of Marketing at PureTech Systems. “These markets are looking for new types of video targets (i.e., street signs, safety scenarios, etc.) and they desire the results of video analysis to aid in navigation and collision avoidance.” Olson explained that video analytics has evolved to address these new markets with the ability to compensate for the movement of the platform itself (i.e., car, drone, boat, etc.), learn new types of targets, integrate with control systems, and evaluate larger video streams with smaller, more affordable software and hardware solutions.

Crowd management is another major trend influencing VCA development, according to Rustom Kanga, CEO of iOmnicient. “Customers are finally realizing that their highest risk is in very crowded environments, and they are now asking for systems that work in such environments,” he said.

In fact, MarketsandMarkets expects counting and crowd management applications to grow at the highest rate from 2016 to 2021. They attribute this growth to how the application could help end users measure the flow of people at specific entry and exit points within a facility, door or building in real time or periodic reporting. Furthermore, the application’s ability to intelligently identify bottleneck areas and alert personnel with the location enables end users to manage the crowd and analyze the behavior and other activities of people and vehicle in real time from remote locations. This not only helps to maximize security, but also aids with decision making.


Better cameras and affordable sensors

Trends aside, advancements in technology and more affordable sensors have also played a major role in VCA development.

Improved camera resolution now provides sharper, more detailed images for VCA. “This increase in resolution avoids the field of view loss caused by zooming into a scene, as well as complications that can arise due to long focal lengths,” Olson explained. These high-resolution images help the accuracy of VCA, particularly in critical environments.

Not only has camera resolution gotten better, cameras have also become smarter and more affordable, both of which contribute to VCA adoption and improved accuracy.

Olson also pointed out how video analytics has become more collaborative over the past few years. “Rather than performing all of its intelligence exclusively based on video data, it has evolved to take advantage of other sensors,” he said.

A more collaborative effort has been in part made possible due to the availability of more affordable sensors. New low-cost position sensor technology provides extremely accurate pitch, roll and yaw position data, according to Olson. This data is then utilized by VCA to allow rapid camera setup for accurate video analysis and exact camera pointing for very long range detection.

More affordable active sensors, such as lidar, radar and thermal cameras, are also making it possible for cost-sensitive industries to use more advanced sensor collaboration via video analysis, Olson said. “These solutions, once restricted to military use, can now be affordably deployed for the protection of all types of facilities requiring this increased level of performance.”


Deep learning and VCA

Deep learning, a branch of machine learning, is also playing a role in the evolution of video analytics. As a matter of fact, the development of more sophisticated computer vision and deep learning algorithms was named a key growth driver for video analytics by market research company Tractica. “When combined with video analytics, deep learning can filter out noise and provide additional insights into the scene, such as improved target classification,” Olson explained. “This becomes extremely useful to more accurately identify targets, especially in complicated scenes and high-resolution video feeds.”

Despite the benefits of deep learning, Kanga noted that these technologies perform heavy computing and may require specialized hardware. As a result, iOmnicient uses a hybrid of heuristic and neural network artificial intelligence technologies instead. According to Kanga, these technologies can achieve the same results as deep learning, but are much less computing intensive and can operate on commercial off-the-shelf hardware.


Working with IoT

Last year ABI Research named the use of video analytics as a business intelligence tool as one of the major trends in video surveillance depicting its convergence with the Internet of Things (IoT).


While retail is the most obvious and biggest user of business intelligence, VCA also works with IoT in other ways. For instance, the use of wearable cameras for law enforcement.


“A system comprising video analytics, wearable body cameras, a vehicle camera, and real-time audio can work together to aid law enforcement to recreate the scene through metadata for real-time alarms, easy post-event analysis and sharing,” Olson said. Such a system, could create a map-based display of the officer and other people in the scene, align audio to specific people, collect vehicle details, and even understand the types of actions occurring in the scene. This information is then available in real time and stored for forensic use or protection support.

Interestingly, a recent article by McKinsey & Company, a worldwide management consulting firm, took a look at how video analytics is shaping IoT, instead of the other way around.

According to the article, published in December 2016, IoT applications offer more value when incorporated with video analytics. This is because video analytics allows them to consider a wider range of inputs and make more sophisticated decisions. As an example, the article refers to how some IoT applications use beacons that transmit location data each time they connect with a consumer smartphone in a store. This data without video analytics would only help retailers track the number of visitors; however, with video analytics, more detailed demographic information, such as the genders and ages of the shoppers would also be provided.

Although the current market for IoT video analytics applications is small, McKinsey & Company believes that there will be large opportunity for it in the coming five to 10 years. This type of application will also become more valuable to a wider range of use cases. As a result, they believe that video analytics could be one of the most important growth drivers for IoT.

Beyond safety and security

In order to extract the most from a VCA investment, end users are looking for their analytics to help improve operational efficiency and provide business intelligence. Aside from its uses in the retail sphere, VCA is also being used beyond security applications to count objects for operation management and detect skimming devices at banks.

For example, VCA is being used to count vehicles to aid in efficiency and revenue generation at large parking facilities and venues. Olson explained, for these businesses, understanding the number of available parking spots and efficiently routing drivers to these locations is key to customer satisfaction and maximizing revenue. “The use of video has proven to be very accurate and affordable, given the use of off-the-shelf video cameras and the ability to deploy through non-intrusive installation,” he said.

How banks use VCA to monitor skimming devices on ATMs is another example. These devices are transparent and almost invisible, Kanga explained. “They result in millions of dollar of ATM fraud each year. The system can detect such devices. Further the system can capture the face of the person who placed it there and raise an alarm if he turns up at any other ATM.”

What’s next for VCA

The number of applications for VCA will only continue to grow as its features and capabilities advance. Especially now, when consumers expect security products to incorporate the latest in consumer technologies, VCA vendors must keep up with trends in order to stay competitive. Also, as more use cases become available, what VCA could be capable of will only be limited by what end users have yet to think of.


Product Adopted:
Software
Share to:
Comments ( 0 )