The developer of a British surveillance system claims its set-up will do something car alarms and The Club can't -- ferret out car thieves before they strike.
By monitoring the strides of people as they walk through parking lots, the controversial system can reportedly detect the difference between people who are searching for their own cars and those who have an intent to steal, the Electronic Telegraph reports.
Researchers at Leeds University are developing people-tracking software based on a form of pattern recognition they believe will outdo current tracking software. At present, this type of software is thrown off the trail by random visual events such as shadows, poor lighting, and reflections from puddles in parking lots.
The scientists have bypassed this problem by using an algorithm taught to recognize the way people ordinarily walk through parking lots to build up a normal model of movement. From this model, the system can predict vehicle and people movements. Motions falling outside the learned patterns will be considered to be those of a potential car thief.
This algorithm, along with vehicle-tracking software developed by Reading University, will be part of a "smart camera" system. Researchers predict it will one day be used to catch shoplifters, bank robbers, and even terrorists.
