Slashdot Mirror
Airport Video Surveillance Goes Hi-Tech
conq writes "BusinessWeek has a piece on new tech used in the airport of Helsinki to monitor behavior and alert people when predefined situations arise. From the article: "The system can alert staff to events which may need further investigation without the need for every camera to be observed by staff. For example, suspect packages or vehicles left unattended will be flagged up and staff alerted. Similarly if the system detects queues growing beyond a pre-defined length in the security zone staff will be alerted of the need to open another lane""
I design security systems just like that one and I know that the video analysis software is not yet good enough to pick out "suspect packages". Sadly, relying on the passengers in the airport or commuters in the rail/subway stations is still the most-effective method of identifying suspect/left packages.
I know of one security system where it just has a picture of a room and then when the camera is on it just looks for things that don't match the original, if those areas of non-matching do not move for more than X amount of time the system then draws a yellow outline around the object, if it still does not move in another X amount of time, the outline becomes red. The system is pretty good and adapts to furniture/plants being moved around by staff. Myself and others even joked about the cheese factor of using colours like that, but it was/is a very robust method of surveillance.
This system was in place in a casino, all the way back in 1998.
Why are these things being labled as new technolgy?
I'd be more impressed if they had a system that could scan and tag/match all faces/voices that it sees and hears in an airport.
False positives tend to be quickly ignored if they happen to often. For instance, in the US a bunch of CIA operatives just recently decided to test US border security to see if the border guards would stop a truck filled with dirty bomb material. Not many people know this, but the US has radiation detectors along its border. The CIA drove a truck with dirty bomb material in from Mexico using forged papers. They set off radiation alarms, had their papers briefly checked, and were sent through.
Now, why did this happen? You would think that setting off a radiation alarm would result in papers getting a good checking over such that their forged nature would be revealed. As it turns out, the radiation detectors along the border go off all the time. They go off roughly a 50-100 times a day at some points. The border guards were so used to the detectors going off that they stopped paying any attention to them.
If a system gives false positives, the danger isn't that the security people will spend too much time examining false positives, it is that they will ignore all positives. I have a feeling that this security system is borderline useless because of this. I find it very hard to believe that this security system can tell the difference between a piece of luggage that is sitting all by itself in a suspicious manner and a piece of luggage resting at a guy's feet who has just put it down so his hands are free to read a newspaper.