Traditionally, the use of video technology in perimeter protection has played a secondary role. “In the past, outdoor video was forensic,” explains John Romanowich, president and CEO of SightLogix, a Princeton, N.J.-based manufacturer of video cameras with built-in analytics designed especially for perimeter protection.
In the traditional set-up, which some installers still use today, primary protection is provided by motion detectors, seismic sensors, photoelectric beams or like sensors. But because those sensors may have difficulty distinguishing between a human intruder and an animal, for instance, one or more cameras are often installed to provide a visual record of what actually occurred. Often the visual evidence is reviewed after the fact — and as Romanowich explained, more often than not the video footage does not clearly reveal what happened.
Video equipment from companies such as SightLogix; Melville, N.Y.-based Verint Systems; and Bedford, Mass.-based VideoIQ are designed to be used for perimeter protection without the need for additional sensors or to help provide extra high security when used in combination with such sensors. Vendors say these analytic systems can reduce false alarms dramatically while also providing prompter notification and better evidence when human intrusions do occur.
Video analytic-based perimeter protection isn’t inexpensive. Andrew Bowman, vice president of engineering for Sanford, Fla.-based integrator Site Secure, estimates the installed price of a single camera to be about $15,000 to $20,000.
At that price, these systems today are most often used to protect extra-high-security areas such as electrical substations, cellular antenna systems, airports, seaports, train tracks and refineries. The government requires perimeter protection for some of these facilities to help protect critical resources and infrastructure (see related story, “Regulatory Requirements” below). Although those mandates do not specify the type of equipment to use, decision-makers may prefer video analytics because of its ability to minimize false alarms.
Video-based perimeter protection is particularly well-suited to installations that do not lend themselves to fence-based protection, such as seaports. Port authorities may use a fence to secure the land side of their sites, while video analytics is used on the water side, where cameras can be programmed to alert when vessels of a certain size approach.
Railway sides are another popular application. SightLogix clients in the rail industry have been able to catch people putting graffiti on freight cars. Graffiti is a serious crime for these clients, who must pay about $5,000 to refurbish just one car, and who also like to avoid what Romanowich calls “the brand problem.”
Car dealers also have a lot of valuable property to protect, so more and more of them are using video-based perimeter protection. Doug Marman, chief technology officer and vice president for VideoIQ, recalls one car dealer client that was able to take down the tall fences that previously surrounded its lot after installing video-based perimeter equipment. Cameras mounted on the showroom building facing out into the lot, in combination with two-way audio that enabled guards to warn people away when the lot was closed, were sufficient to secure the site. And, without the fence, Marman says, sales increased because potential buyers were able to more easily view the inventory.
Despite their high price tag, video analytics can be less expensive to install than other forms of high-security perimeter protection in certain applications, because they can eliminate the need to dig trenches for buried cable or to install a fence.
“Trenching to the fence is the most expensive part of the installation,” Marman relates. As an alternative, it often may be possible to mount cameras on the side of a building — a choice that Marman thinks is appropriate about 80 percent to 90 percent of the time. And even when cameras must be pole-mounted, it may be possible to use wireless communications.
End-user organizations also may be able to justify the expense of a video analytic installation because it can reduce the amount of human guards needed. “You’re taking costs out of traditional badged employees that would be moving across the perimeter and properties,” notes Verint Systems senior vice president and general manager Stephen Foley.
Installation Planning
Vince Nussbaumer, security division manager for Houston-based security integrator BL Technology, estimates that about half of his company’s perimeter protection installations involve upgrades to existing security systems, while about half are installed where there previously was no such system. BL Technology originally focused on wastewater facilities, but since has branched out into other types of clients.
If a site already has analog cameras, it may be possible to minimize the cost of the upgrade by adding analytics to those cameras. Nussbaumer says BL Technology has had success using ADPRO equipment from Norwell, Mass.-based manufacturer Xtralis to add analytics to analog cameras. For brand new installations, the company uses SightLogix cameras which have built-in analytics. Another option that some integrators may use for converting analog cameras is to add analytics along with an encoder that converts the video stream to IP. Encoders with built-in analytics also are available, as are cameras with built-in storage that can minimize the amount of video sent across the network.
For jobs where cameras do not already exist or will not be reused, several manufacturers offer software tools that can help determine how many cameras will be needed. To use these tools, typically based on Google maps, installation planners input the expected latitude and longitude of each camera, along with the anticipated camera mounting height. The software then displays a pictorial map of the installation, with the location of each camera and coverage area superimposed on top.
Such mapping tools can be helpful in compensating for the dead zones that comprise as much as one-fifth of the area within the coverage range, including the area directly underneath the camera and extending slightly in front of it, Romanowich notes. He recommends installing an inner tier of cameras to back up the ones at the outer edges of the site. “You can take the camera that is in behind and move it forward a bit so it watches the back of the other camera,” he explains.
If a camera must be pole-mounted, installers may opt to use wireless transmission. Typically the cameras are programmed to only transmit video in the event of an alarm condition, minimizing the amount of bandwidth required. As with any wireless installation, ensuring that each transceiver has line-of-site to another transceiver is critical. When wireless communications will be used, Foley advises integrators to test the quality of the wireless link early in the installation process.
Integrators also should consider whether there is sufficient light for a day/night camera or whether an infrared light source might be needed. “If there are no streetlights, you know you will have to add infrared,” Marman observes.
Another consideration is weather conditions. Cameras may not capture good enough quality video in heavy rain or dust conditions. And for high-security installations, it may be a good idea — or even a regulatory requirement — to use another technology as a back up.
The most difficult aspects of planning video analytic-based perimeter protection may not be technological, but psychological. Bowman’s biggest concern is managing customer expectations about what a system can do — and he believes quality varies significantly from one manufacturer to another. His advice is to borrow equipment from at least two different vendors and install it at the customer’s site on a trial basis so the customer can compare systems and know the level of performance to expect. “Most manufacturers are willing to loan the equipment but they’ll want it back in short order,” Bowman says.
What Else?
Marman and Romanowich say their companies work closely with the client for the first installation and after that, most integrators are able to handle future installations on their own. Although there are many parameters to program, most software is menu-driven, and as Marman puts it, “If you know how to do video motion detection, the set up is easy.”
Nevertheless, one element of video perimeter protection is considerably more complex than the rest — and that’s the command-and-control center, which is manned 24/7 and is a critical element of virtually every perimeter protection system. As Krag Reinertson, business development manager for Herndon, Va.-based integrator GTSI notes, video analytics often is just one element feeding into broader scale “situation awareness” based on the physical security information management (PSIM) standard. Video perimeter protection can “tie in very nicely” to this type of system, Reinertson says, but such systems require integrating the perimeter system with other systems such as access control and sensors.
Typically these installations have video walls with multiple screens that display the status of various sub-systems. The systems correlate alerts and other data from these subsystems, automatically displaying what they determine to be the most relevant video images to command center personnel. In the event of multiple intruders, these systems also may have the ability to plot the position of each intruder on a map and automatically display video associated with each intruder, calling upon one camera after another as needed.
Another approach, available from Ann Arbor, Mich.-based manufacturer Pixel Velocity, “stitches together” images from multiple cameras so that control center personnel have a more comprehensive view of the protected area. Pixel Velocity’s technology enables personnel to view as many as 20 cameras that have been linked together into a single image, notes Jason Slocum, Pixel Velocity director of national and global accounts.
When all of the elements of video analytic-based perimeter protection are taken into account, it’s not surprising to learn that the site acceptance process can take several days. “The customer could have hundreds of tests to do,” Romanowich describes.
Typically customers conduct acceptance tests with the systems integrator — and sometimes a representative from the equipment manufacturer — at their side. A thorough acceptance test should include a walk-test of each camera; the best way to walk-test the camera is to walk towards it, because that is the most difficult type of motion for the camera to detect, Romanowich advises.
“They should also make sure the overlap is properly covered and that in the far field, the equipment does what it’s expected to do,” he adds.
Regulatory Requirements
For some end-user organizations, perimeter security is not optional — it’s a government-mandated requirement. Typically these mandates say the organization must “deter and detect” intruders.
The Department of Homeland Security (DHS) has issued requirements for airports, seaports and other facilities, including water treatment plants, which are covered by Chemical Facility Anti-Terrorism Standards (CFATS). The Federal Energy Regulatory Commission (FERC) and North American Electric Reliability Corporation (NERC) have established requirements for power plants, electrical substations, etc.
For more information, visit: Department of Homeland Security (www.dhs.gov), Federal Energy Regulatory Commission (www.ferc.gov) and North American Electric Reliability Corporation (ww.nerc.com).
Video Analytics as an App
In the article “Applications… Security Style,” (published March 2010, page 163), SDM reported on a new approach to gaining analytic capabilities without purchasing it as a built-in feature of the camera. This approach was announced recently by Axis Communications, Chelmsford, Mass.
With the launch of the AXIS Camera Application Platform, it is possible for any company to develop downloadable applications for Axis’ range of network cameras and video encoders. More importantly, it will allow integrators and dealers to place custom analytics inside Axis’ cameras, harnessing the processing power that is available there.
“We see great interest for intelligent video applications that provide the products with capabilities based on, for instance, recognition, tracking, detection and counting,” said Ray Mauritsson, chief executive officer of Axis Communications.
Fredrik Nilsson, general manager of Axis Communications, explains further: “We are not the expert analytics company, but we are experts in partnering and will open up the cameras’ processing power and access to uncompressed video to our video analytics partners in order to provide the best possible best-of-breed solution for the integrator and end users.”
