Review Course on Video Surveillance

Most of us tend to date the onset of shooting violence in schools from the Columbine massacre in 1999. In fact, gun violence has plagued K-12 schools as well as colleges and universities for decades. Wikipedia lists 103 incidents of gun violence in schools since 1966. Of those, 67 incidents occurred in K-12 schools.

Columbine did, however, underline the need to find better ways to use video technology. “Since Columbine, many K-12 schools have added video,” says Paul Timm, PSP, president of RETA Security, Inc., a consulting firm specializing in school security. “I think it is rare today for a school to have no video capabilities.”

Columbine also ignited an effort to improve video surveillance capabilities. While surveillance video of the massacre helped officials piece together what happened after the fact, the shear depth of the tragedy raised questions as to how video (and other security technologies for that matter) might have been used to give officials a better understanding of what was going on inside the school, perhaps enabling them to reduce the death toll.

Among the innovations developed in answer to that question are Internet protocol (IP) digital cameras that plug directly into computer networks, making it easy to add cameras and possible to monitor video from remote locations.

In addition, new, high-resolution megapixel cameras produce clear, detailed full-color images that are easier to grasp than the grainy black and white images most often associated with surveillance video.

Then there are video analytics. Although early video analytics technology failed to live up to its advertising, today’s smart video systems perform better and are finding effective applications in schools.

Brought to You by the Network
CameraWATCH Technologies, a video technology integrator in Jackson, Miss., is installing more IP cameras in schools these days. “A district south of Jackson recently installed new IP cameras in all five of its schools,” says Jim Walker, vice president with CameraWATCH. “The schools already had an analog camera system. The IP cameras expanded coverage.”

Older analog cameras require expensive coaxial cabling to carry the video as well as power sources. IP cameras connect easily to existing Ethernet network cabling. In addition, an IP camera feature called power over Ethernet (POE) provides electricity to power the cameras and eliminates the need for special power supply installations.

IP cameras have given rise to some innovative applications. “One superintendent ordered one IP camera installed on a tripod for each school in the district,” says Walker. “If behavior problems arose in a classroom, the teacher brought in the camera and plugged it into the network outlet and let it run, telling the class that the superintendent and principal were going to watch the class for a while.

“Under the camera’s watchful eye, a number of discipline problems disappeared.”

IP cameras necessarily involve IT directors in the security mix, since video data will compete for bandwidth with the network’s other data responsibilities.

David Massaro, technology services coordinator, with the Redlands Unified School District in Redlands, Calif., has been managing the installation of IP cameras throughout Redlands’ 23 elementary, middle and high schools since 2005.

The district began the video rollout at Redlands High School, deploying 18 IP cameras supplied by San Juan Capistrano, Calif.-based IQinVision. The success of that installation led to a second IP camera installation at Citrus Valley High School. There, the district installed more about 30 IQinVision dome cameras inside the school and nearly 60 exterior cameras.

“Now we’re putting cameras in all 15 elementary schools, four middle schools, the third high school and a continuation school,” Massaro says. “Because they are all IP cameras, they are POE, and we don’t need to install coaxial cabling — we just plug into a data jack.”

Security officers monitor the cameras by using a browser — from the district security center or home, if necessary — to access a video recorder and camera management system supplied by On-Net Surveillance Systems, Inc. of Pearl River, N.Y.

Massaro has configured the system to record cameras from five to 15 frames per second and to store video for 16 days, which is long enough for an incident to be reported and checked out.

“It’s a very intuitive system,” Massaro says. “It shows a layout of the school, and you click on a camera icon to view video from that camera. It takes five minutes for a principal or security officer to learn to use it.”

More Pixels, More Picture
In addition to operating over the school district’s data network, the Redlands Unified video cameras feature high-resolution megapixel cameras.

The word pixel comes from the term “picture element,” and it describes the smallest unit of a digital image. A one-megapixel image has one million picture elements or pixels.

The term pixel also describes single sensor elements on the sensor of a digital camera. A one-megapixel camera has one million sensor elements. 

A one-megapixel camera costs twice as much as a standard definition camera.

“But it provides four times the resolution,” says Paul Bodell, executive vice president of global business development with IQinVision.  “A standard definition camera can cover an area eight feet wide by eight feet deep and provide good quality images. Four standard definition cameras can cover an area 16 feet wide by 16 feet deep and provide good quality images. A single one-megapixel camera could replace all four of those standard definition cameras, for half the cost.”

Generally, the interior cameras at Citrus Valley High School provide 1.3 megapixels of resolution, while the exterior cameras provide between three and five megapixels of resolution. “The multi-megapixel cameras provide a big field of view,” Massaro says. “You can also zoom in on sections of the images without pixelating them.”

Sergeant Dan Kivett manages the security staff at Citrus Valley. Kivett says that the high-resolution video has helped deal with minor thefts, scuffles and even serious fights. “We can recognize faces on the video,” he adds. “Students are quickly getting to know this, and it is becoming a powerful deterrent. Word of mouth is probably our best tool, and it travels fast.”

Megapixel cameras offer other benefits as well, says Bodell. Students quickly figure out how to time cameras that pan, tilt and zoom to cover a larger area. Megapixel cameras don’t need to pan, tilt or zoom. With wide-angle lenses, fixed megapixel cameras can capture everything. Timing the movement of a panning camera no longer works.

Megapixel camera power makes it possible to invent new kinds of camera applications. “One company recently introduced a hall-view camera,” says Walker. “Instead of use the basic high-resolution format of 16 by 9, you can reverse it and stand the picture on end and shoot all the way down a hallway.”

High-resolution combined with a wide, deep field of view and eliminating the need for pan, tilt and zoom cameras also help to make video analytics technology more effective.

Monitoring Video With Technology
Video analytics technology got off to a bad start. Promoted as a technology capable of monitoring video surveillance cameras and detecting certain programmed scenarios — fights, chases and, among other things, people loitering or cars parked in a particular area for too long. In all, analytics providers claimed to have about two-dozen scenarios that an IP surveillance system could monitor for and alarm on.

Users complained, however, about too many false positives. Users also found that they could not use pan-tilt-zoom controls on cameras equipped with analytics because the motion would confuse the analytics and set off false alarms. Finally, some of the scenarios programmed into analytic systems were too ambitious. The programming wasn’t good enough and the processing power of the computer chips wasn’t powerful enough.

“Today, the algorithms that program the scenarios into video analytics systems have improved, along with the processing power of the computer chips,” says Fredrik Nilsson, general manager with Boston-based Axis Communications, North America, a camera supplier that includes video analytics in its cameras. “In addition, megapixel cameras provide more detail to the analytics. Together, these three things are making video analytics work better.”

Finally, the travails of earlier systems have scaled back expectations. Today, scenarios are more modest. “Analytics are used to detect people moving around a building after hours or at night, which no one should be on the property,” continues Nilsson. “Another application detects people congregating into groups, which could indicate a fight.”

In addition, applications can be created to attack problems at individual schools. For instance, Redlands Unified School District monitors its own fixed megapixel cameras. But the Redlands, Calif., police have their own pan-tilt-zoom cameras on campus too. Massaro is considering an analytics system that would detect an issue on the school district cameras and then send an alarm to the police department’s pan-tilt-zoom cameras and cause them to zero in on the event.

Clearly, the use of video surveillance cameras in K-12 school has made great strides since the Columbine massacre. Today’s systems use computer networks to install more cameras and to monitor them from anywhere. They use megapixel cameras to record more detailed scenes. And they are beginning to find value in video analytics technology, which can help ensure that the camera system is being adequately monitored.