Surveillance Systems Getting Smarter
by Tom Olago
Surveillance cameras are now able to go “over and above the call of duty”, so to speak, where all they could do traditionally was to capture and transmit still and video images. They have now been advanced to a point where they are getting better at picking you out of a crowd and recording where you’ve been, what you are doing now, and where you are going after you are through. ABC news recently reported on these developments, quoting Jenq-Neng Hwang, electrical engineering professor at the University of Washington and lead author of a new study.
According to the report, Hwang’s team has developed technology that will allow multiple cameras to follow an individual as he or she moves through a crowd, switching seamlessly from one camera to the next as the target moves from one field of view to another. It can do that, even if the target disappears for a while and then reappears in a different area.
His experiments linked several cameras in a network and they were able to follow one target all over the Seattle campus. The cameras were able to “talk” to each other, so the target could be highlighted, making it easy for the next camera to pick up the chase. They could do that, without human supervision, even if the target’s face could not be seen. This would work anywhere, as long as the cameras are part of a network and can upload data to the cloud.
This system is still in the development stage, but Hwang pointed out that monitoring systems are already installed in many buildings and institutions, so it’s not a giant leap to move from a broad surveillance network to a smart system that can focus on a single target.
The report also mentions some developments that could result in potential improvements to existing smart surveillance capacity and techniques:
• Researchers at the University of California, Riverside, are training cameras to pick out suspicious behavior, like loading a gun or putting an object in a pocket so the clerks in a store can’t see it.
- Medical researchers at the University of Oxford want to see us all wearing cameras that will tell our doctor whether we are taking our pills and getting enough exercise.
- Scientists at Carnegie Mellon University in Pittsburgh have developed a system that will monitor every activity in a nursing home. Their system is also capable of tracking a single individual, based on color of clothing, facial recognition, or other factors, and it is beyond the experimental stage. It has actually been installed in a nursing home.
- Banks are able to monitor traders who may attempt to conduct illogical or unethical trade practices. The office surveillance tools in use spy on trading relationships and behaviors. Mail Online reported: “Stung by billion-dollar fines for malpractice on their trading floors, the world’s big banks are using ‘fuzzy logic’ tools such as relationship mapping and behavioral analytics to ‘read the minds’ of would-be cheats among their traders.
Indications are that these surveillance systems are being developed to a point where they can minimize the need for monitoring personnel to virtually none, except perhaps for required oversight of equipment functionality. The days when human beings will be required to view and interpret still picture images and video footage are clearly numbered.
According to Seagate.com, traditionally, the task of surveillance video review has fallen to trained security personnel. Capable of monitoring a modest number of incoming video streams, such personnel become progressively less effective as the volume of video data grows and overloads the ability of the human eye/brain to process visual details. Exacerbating the problem is sheer fatigue, with long shift hours degrading the monitoring abilities of security staff still further. Adding more security personnel is a very costly option that, while reducing the number of video streams each security staff member must review, still fails to eliminate the inherent disadvantages (limited attention span, interruptions and distractions, fatigue) that accompany human monitoring of video data.
IVS (Intelligent Video Surveillance Systems) software overcomes such limitations, enabling video surveillance equipment manufacturers and system integrators to create intelligent video solutions that see and process visual information similarly to humans. For example, such video analytics systems can distinguish between a person and a car. These systems can be programmed to track only objects identified as human and send an alert when the subject violates predefined rules, such as climbing over a wall.
Seagate lists the following advantages of Intelligent Video Surveillance Systems (IVS’s):
1. More Cost-Effective: Conventional video surveillance environments require security personnel to spend many hours watching live or recorded video to analyze/identify suspicious events. By contrast, IVS systems can scan many thousands of hours of video data without human intervention. Should the IVS system encounter questionable activity, it can automatically notify security personnel for further investigation. By increasing the number of video streams that can be handled by each security team member, IVS systems can significantly cut personnel costs.
- More Accurate: A variety of tests have shown that humans lose anywhere from 50% to 90% of their visual perceptibility after 20 minutes of continuous video monitoring. The more video streams a person is required to monitor in a given period, the sooner impairments in visual perception manifest themselves. By contrast, IVS systems are immune to the fatigue, distractions and memory lapses that plague human beings. Live and recorded video review and analysis by IVS systems continuously scour video images for user-selected behaviors and patterns (for example, intruders, license plate numbers, unattended baggage), while ignoring the extraneous data that leads human monitors to generate false alarms.
- More Scalable: Security environments are dynamic, changing both in size and character as a business grows. Thus any video surveillance solution must be able to scale and adapt as needs dictate. With a conventional security system, that means hiring and training more personnel to monitor any additional camera streams – a costly and time-consuming process.
Is it any wonder then that such smart surveillance systems are said to be the next level in cutting edge surveillance? Privacy advocates seem to be fighting from increasingly disadvantaged positions, as smart surveillance seems to score more points relative to their shortcomings. And as the ABC news report aptly concluded: “So big brother is already watching. But he’s getting a lot more efficient.”