Analytics Answers
Analytics Answers By Buddy Flerl Video analytics has the potential to tremendously impact the security industry. By using image processing algorithms to detect, track and analyze specific activities captured in video streams, video analytics software can accurately identify threats in real-time the moment they begin to unfold and alert personnel to take immediate action.
In many installations, video analytics software protects investments in CCTV systems by means of operating on a centralized server that communicates with existing cameras through existing networks. As with any young technology, though, not all video analytics software is created equal. A solution that’s perfect for one installation may have significant issues in another.
When investigating any video analytics solution, the following 10 questions will help security contractors determine if the proposed solution is thoroughly suited to its target environment, and whether it will succeed in meeting the security challenges it is intended to address.
1. How scalable is the system?
Security systems typically grow over time to meet growing needs for coverage. This mandates a system be truly scalable and “future proof.” The first consideration for scalability is the number of cameras a single server can support. For example, if each server is limited to eight cameras, adding a ninth camera requires adding a new server.
The added server will be grossly underutilized until the eventual addition of more cameras. Once seven more cameras are added, the process starts again, with another new server needed at the next camera addition.
The second consideration for scalability is the video analytics architecture. Several video analytics systems allow you to add new cameras, one at a time, to expand a system from a few to hundreds or even thousands of cameras. The supporting hardware may be less accommodating, however, depending on the video analytics architecture.
Video analytics typically takes one of two forms: server-based or edge-based. Each has its own unique strengths. Server-based solutions are more flexiblethan edge-based but create an additional burden on the network. High quality compressed video needs to be sent to these servers in order to provide high quality analytics. As these systems scale, this burden increases. In effect, they may actually double the network utilization when doing analytics plus recording and viewing, vs. just recording and viewing.
2. Is the analytics system compatible with existing network and video equipment?
All video analytics software must interface with a network and with video cameras to deliver its magic. The software typically resides on one or more servers and sometimes in the security cameras (depending on the implementation), with the network handling server-camera communications.
For existing video systems it will be necessary to ensure the analytics software being considered is compatible with the network and video equipment already in place. There’s really no reason to scrap an existing investment in cameras and networking to accommodate new software. And to ensure compatibility with any related equipment that might be installed in the future, it’s best to go with an open, standards-based system.
3. Does the system provide real-time monitoring and alerts?
The passive nature of a video surveillance system is an inherent deficiency; its automation is limited to video delivery.
To be useful in addressing unfolding threats, a video surveillance system must be continuously monitored by attentive security personnel who can issue an alert or take action themselves. These systems often serve primarily as evidence recorders, offering video of a security breach that can later be examined to identify event circumstances and perpetrators.
Improving security with real-time monitoring and alerts is the main reason for adopting video analytics. You’ll want the system to automatically alert personnel when a potential threat is unfolding; anything less is comparable to traditional video surveillance systems. Make sure the system under consideration is capable of detecting behaviors and scenarios specific to the security challenges of its environment.
4. Can the system run different rules, and a varied number of rules, within each camera?
Unless installed in an unusually simple environment, the video analytics system should be capable of detecting specific types of security threats in one location, and entirely different types of threats in another setting.
A retail operator, for example, might be concerned with simultaneously monitoring multiple points on the property, such as:
- A crowded parking lot where threats to shoppers and their vehicles exist
- Loitering near expensive merchandise or for anyone reaching over an unattended counter
- A loading area for activity during nondelivery hours or for objects placed there for later theft.
In this example, each surveillance area requires cameras set up with rules that are unique to that location. Similar scenarios are played out regularly in airports, seaports, transportation facilities, banks, government offices, military installations and virtually any other enterprise-grade organization, public or private.
To meet the need for area-specific security, the video analytics system should allow security contractors to select which rules are installed in each camera, and to install as many rules as are needed to cover any given area’s range of threats.
5. Can the system’s analytics compensate for varying environmental conditions?
Just as most installations have the need for different video analytics rules at different locations, the analysis of video from some cameras may need to account for varying environmental conditions. The camera securing an outdoor parking lot and the camera securing an interior floor are operating in two distinct environments. The video analytics addressing the parking lot must accommodate all types of weather conditions.
Unless the installation is entirely indoors, a video analytics system should be fully operable in both indoor and outdoor settings to account for full daylight, deepening shadows, artificial lighting or glaring sunlight, plus adverse and varying weather conditions. The software should compensate for background interference such as moving trees. It should also include filters for shadows cast by moving or stationary objects, variable light levels and such random weather factors as clouds, rain, snow, ice and wind.
6. Can the system truly achieve 99-percent accuracy?
Claiming that a video analytics system delivers 99-percent accuracy is like claiming that an Enzo Ferrari can run at speeds above 220 miles per hour. Whether it actually does depends on the circumstances.
While some current video analytics systems are far more accurate than the first wave of solutions to hit the market, accuracy claims for video analytics warrant evaluation. Under what conditions were the measurements taken? Was it an indoor setup with controlled lighting and modest rules, or outdoors at twilight on a windy, rainy day with highly specialized rules? And if something in between, what were the circumstances?
The real question is can the system deliver the highest accuracy for the conditions under which it will be deployed?
7. What are the economics of the system?
The true cost of scaling a video analytics system depends on the camera-to-server ratio, and server-based vs. edge-based issues, discussed above.
With any video analytics system the basic expansion costs begin wit
If you enjoyed this article and want to receive more valuable industry content like this, click here to sign up for our FREE digital newsletters!
Related Content
Security Is Our Business, Too
For professionals who recommend, buy and install all types of electronic security equipment, a free subscription to Commercial Integrator + Security Sales & Integration is like having a consultant on call. You’ll find an ideal balance of technology and business coverage, with installation tips and techniques for products and updates on how to add to your bottom line.
A FREE subscription to the top resource for security and integration industry will prove to be invaluable.
