To put it mildly, the Qingzang railway is a triumph of human engineering over nature. But given its location in the frozen “Rooftop of the World,” and continuing political unrest in Tibet in response to Chinese rule, it makes sense for the Qingzang’s managers to keep an eye on the line from end to end.
This is where NICE Systems comes in. NICE offers a video analytics system that allows Qingzang railway personnel to conduct remote video surveillance on the line, from company headquarters in Xining, China. More importantly, the thousands of weather-hardened NICE edge devices deployed along the line are able to decide whether the railway’s monitoring center needs to see their feeds or not, based on what is being detected within their fields of view.
“Our video analytics-enabled camera system keeps Qingzang staff apprised of what needs to be responded to, and what doesn’t,” says Moti Shabtai, executive vice president of sales for NICE Systems in the Americas. “In this harsh environment, you need this kind of monitoring to ensure the safety of passengers, crews and rolling stock.”
CAMERA PLACEMENT, HIMALAYAN STYLE
A track 1,215 miles long can eat up a lot of surveillance cameras. This is why the Qingzang railway tried to balance coverage against cost, by locating poles at one-mile distances. The poles were installed by the railway itself, along the snowy tracksides. The cameras—typically five on each pole—were mounted by NICE Systems contractors with the aim of providing both wide angle and telescopic views from each location.
“Each of the edge devices is installed inside a weatherproof case with an electrically-powered heater and cooling fan,” says Shabtai. “The camera lenses are all fixed length, since the system is not configured to support remote pan/tilt/zoom functions. Combined, each surveillance location is designed to give the security monitor a clear view down both directions of the track. Properly done, the result is seamless coverage along the whole line.”
Typically, each set of SD resolution cameras is mounted 24 to 36 feet above the ground, to provide the longest possible view and protection from vandals.
The cameras themselves come with digital recorders and video analytics encoders. The recorders capture the video in a buffer, with any video of interest transferred electronically to another location for long-term storage. (These transfers are triggered by human operators. Otherwise, the buffer automatically overwrites itself.) The video analytics encoders are used to analyze the video detected on the camera’s CCD to determine whether the Xining security center needs to be alerted, and sent the video.
It is the pixel-by-pixel analysis that counts. Essentially, the camera’s video analytics software is programmed to accept a certain distribution of pixels; such as that of an empty rail line scene where nothing is moving in the image. User-defined rules within the software also tell the system what kinds of movement are acceptable; such as blowing snow, or cows moving in the distance. But a shape that appears to be a man moving along the tracks, or fallen debris along the line—that isn’t acceptable. When this happens, the video analytics software sends an alert down the line to the Xining monitoring center. The human monitor then accesses the camera channel remotely, to see what has triggered the alert.
“By having the cameras do the initial analysis and only send alerts when their security parameters are violated, we vastly reduce the size of the network that has to carry this traffic,” Shabtai says. “We also substantially reduce the strain on the human monitors. Rather than trying to watch thousands of individual feeds, they only have to look at those that may be indicative of trouble.”
FUN TIMES IN SHANGRI-LA
The Qingzang line is not your ordinary railway. Normal railroads don’t have to offer coaches with oxygen supplies and extra UV protection for passengers, when rolling through the highest parts of the Golmud- Lhasa route. Nor do normal railways have to worry about track problems in the middle of the remote Himalayas, or vandalism from disaffected locals.
NICE System’s video security system helps the Qingzang deal with these last two problems. But doing this effectively hasn’t been easy for NICE; especially since its commitment guarantees a very low false alarm rate per camera per day.
“We have had to contend with many challenges in the Himalayas,” Shabtai says. “Our edge devices have to work in very low temperatures and stand up to falling snow. To do this, we have developed electronic filters that can mask out falling snow, and help the video edge devices see better through fog.”
Then there’s the wind: The Himalayas can be a very windy place, especially during blizzards. To deal with this, NICE Systems’ edge devices are programmed to compensate for a shake rate of +/-20 pixels in either direction. This ensures that the video feeds are watchable in virtually all conditions. Where nighttime visibility is an issue, infrared capability is included.
“Our edge device video systems have to take all of this region’s extreme weather into account, and keep on working nonetheless,” Shabtai says. “After all, it isn’t easy to deploy a repair crew during the middle of winter in the Himalayas; even you can get them there by train.”
SURVEILLANCE THAT WORKS
The Qingzang railway was finished on October 12, 2005, with regular service starting the next year. Today, you can buy tickets online at www.chinatibettrain.com. The lowest price fair from Beijing West to Lhasa is $49 for a “hard seat,” $102 for a “hard sleeper” and $158 for a “soft sleeper.” (GV Travel Advice: It’s 2,525 miles from Beijing West to Lhasa is 2,525 miles, so splurge for the soft sleeper.)
Should you ever take this train across the Rooftop of the World, your onboard security will be maintained by NICE Systems’ edge devices and video management system. Likely you’ll see the poles with camera clusters as you roll through the snow. Should trouble occur ahead of your train, it will be these cameras that see it first, and alert the train to slow down or stop before you get there.
“Being part of such an ambitious and astounding video surveillance system has been an enjoyable challenge for NICE,” Moti Shabtai says. “It also proves just how useful video analytics can be, even in the remotest, wildest areas of the world.”