Fredrik Nilsson is the general manager for Axis Communications in North America.
Analog CCTV has served the security market well for decades, but let’s call it what it is: Analog cameras are dumb boxes that can see. The continuing and inevitable shift to IP video presents many more features and functionalities not only for security practitioners, but also newer power users of video surveillance — like business owners, marketing and operation departments, and facilities managers.
That’s because IP cameras are computers with lenses.
Since network video products are computer-like devices, they follow much the same innovation path as our PCs, laptops, smartphones and tablets. Most notably they follow Moore’s Law, named after Intel co-founder Gordon Moore who stated that the number of components on an integrated circuit doubles every two years. This theory was updated by another Intel executive, David House, who claimed the circuitry doubling rate is actually 18 months — meaning processing power doubles while cost and device size remain constant or even decrease.
The exciting part for IP video is that processing power growth in IP cameras and video encoders has actually exceeded Moore’s Law since the IP camera was invented in 1996. This “planned innovation” of Moore’s Law has helped IP video hit significant milestones and lay the groundwork for smarter surveillance today and tomorrow by leveraging the edge.
Defining the Edge
“The edge” is defined differently by different markets and even different people. The networking world’s definition of the edge (e.g. edge routers, edge switches) might be different than a business’ IT manager (e.g. laptops, work stations, smartphones). In the video surveillance world, edge devices are typically those that gather and/or transmit video data to the system’s core — otherwise known as the head-end (VMS system, hybrid DVR, server, etc.).
These devices that sit on the “edge” of the network are IP cameras, video encoders and network attached storage (NAS) devices.
Most surveillance professionals equate “the core” as a video storage function; however, video analytic processing, video management and general data processing are also done at the core. Yet as IP cameras and encoders become more powerful, these functions are being pushed to the edge.
The Edge’s Influence on Video Usability
As processing power of IP video devices increased, the most obvious benefit was that manufacturers were able to deliver improved video quality that ultimately (far) surpassed analog to the HDTV and megapixel video of today. As video resolutions increased, the data created remained usable because of more efficient bandwidth consumption and lower storage requirements at an ever-decreasing cost — thanks again to Moore’s Law.
But the influence of edge processing power doesn’t stop at resolution. As resolution higher than 5 megapixel is limited by the lens; and because other parameters such as low-light performance and Wide Dynamic Range (WDR) are ultimately more important for surveillance, R&D engineers have recently focused on leveraging edge techniques to create breakthrough image usability technologies. Many manufacturers stopped focusing solely on delivering more and more pixels, but instead answered, “How can we make today’s pixels clearer?”
Improved chip technology combined with advanced algorithms inside IP cameras led to finer image focus with technologies like P-iris control, much improved WDR performance to combat challenging lighting conditions, and even enable IP cameras to see color video at night — something analog users could never dream of.
Compression at the Edge
None of these image enhancements would have been usable today, if it wasn’t for the use of efficient compression that enables high-quality video to traverse the network in reasonable packet sizes. These advances in compression standards, along with edge improved processing power for real time compression, have evolved to significantly reduce image file size without adversely affecting visual quality.
H.264 compression (a.k.a. MPEG-4 Part 10) is the de facto standard in IP video surveillance. Without it, HDTV and megapixel quality video wouldn’t be possible. This technology enables manufacturers to improve frame consistency for better efficiency, such as reducing color nuances and improving color fidelity and comparing adjacent images to remove unchanged details between video frames.
Since compression technologies are driven by the consumer electronics world, expect even more efficiency in the near future — namely H.265. Ray Coulombe’s March 2013 Tech Trends column in STE (www.securityinfowatch.com/10878865) points out how H.265 can influence another area some might consider the edge: streaming video to mobile devices in the field.
Storage at the Edge
Video storage at the edge — connecting a camera or encoder to NAS device or using an internal SD card slot — is nothing new. Almost all manufacturers began including SD card slot options by the mid- to late-2000s, but mostly as a redundant storage fail-safe mechanism. However, edge storage hardware came at a premium. Here’s where Moore’s Law left an imprint once again. Only a few years ago a 1GB card cost around $100; today, a 32GB SD card can be found for less than $50, and many IP cameras and video encoders support up to 64GB.
Thus, by the year 2012, the concept of an all-edge storage system was ready. This is where different edge technologies — namely hardware storage and software intelligence — combine to build a decentralized solution where the cameras/encoders act as the recording device. Once installed, not even a computer is required to run these systems, as the VMS runs inside the edge device too.
These edge-recording surveillance solutions are ideal for systems of 16 cameras or less today. They can also be used for temporary installations or areas with limited bandwidth connectivity. And, just as they started, edge recording can of course still act as a failsafe redundant option for larger systems.
As Moore’s Law drives the cost of storage down and storage capacity up, soon we can expect the SDXC (extended capacity) standard to deliver up to and exceeding 2TB of in-camera storage, which could be years’ worth of video inside an IP camera.
The Edge of Genius
When we talk about the future of surveillance, intelligence must take center stage. It’s estimated today that a staggering 99 percent of all recorded surveillance video is deleted before it’s ever seen. Of the one percent of video that is seen by a person, only one percent of that video is viewed live. As the surveillance industry continues to grow and more cameras are installed throughout the world, we are creating even more video channels to monitor and recordings to search.
This is where intelligence plays. Since IP cameras are computers with lenses, they have the ability to use software to improve surveillance efficiency. That means empowering security practitioners, law enforcement personnel and business owners with proactive video alerts.
Like edge storage, video analytics at the edge have been around for quite some time. In 2000, manufacturers introduced video motion detection, the first intelligent algorithm that resided inside a camera. Motion detection is not only used to alert operators to someone or something in the area, but is also an established practice to “record only on motion” and save bandwidth/storage. It was a tremendous first step for IP video intelligence.
Then the analytic and video intelligence market hit a major bump in the road following the Sept. 11 attacks, when the hype and promise surrounding advanced analytics like facial recognition and bag-left-behind came well before the technology was ready. Fortunately for the market, while many users became understandably sour on the success rate of these advanced analytics, effective edge analytics like people counting, cross-line detection and camera tampering alarms emerged from the brief analytic funding utopia.
It also was proof positive that the camera itself can be turned into an intelligent device.
Quality IP cameras and video encoders now set aside a certain amount of memory and CPU cycles dedicated to running applications in-camera. Some manufactures also support open application platforms that allow third-party software developers to write custom apps that can run in-camera — a platform much like a smartphone.
On top of cross-line detection and people/object counting, third-party apps such as license plate recognition, line/queue management and heat mapping can be downloaded to run inside compatible IP video devices. Some leading software providers have even developed a full VMS suite to run at the edge. This is just the beginning.
The term “Big Data” is a very familiar one in the IT world. We have created so much data as a society that companies are now actively searching for the best way to use, organize and leverage all this information. With all that video surveillance data being deleted today without anyone ever seeing it, the IP video market has the same dilemma.
By leveraging the ever increasing processing power of network cameras and exposing young software developers to our industry by partnering with leading universities and IT companies, we can do more at the edge to build an industry of proactive devices. The goal is to help security practitioners process pertinent video they need to see in real-time.
Living on the Edge
With all these surveillance functions moving to the edge — from never before seen image usability technology to recording and intelligence — might it be that one day the edge can accomplish ALL the functionality of the system and no central point is needed?
Those systems are already in existence today, and are especially effective in small systems. Welcome to the future, brought to you by the edge.
Fredrik Nilsson is the general manager for Axis Communications in North America. He is the author of “Intelligent Network Video: Understanding Modern Video Surveillance Systems,” published by CRC Press. His popular “Eye on Video” series can be found on SecurityInfoWatch.com. Request more info about Axis at www.securityinfowatch.com/10212966.