A traffic camera that can identify and count the number of people in a vehicle using infrared imaging has been developed by Loughborough University spin-out Vehicle Occupancy (VOL).
Called Dtect, the multi-spectral imaging system is designed to help cut congestion and pollution on the roads by collecting information about the number of people in a vehicle. It would enable traffic controllers to target lone motorists and make them pay higher congestion charges.
The device is a single, autonomous pod, which carries out imaging, illumination and data processing functions that can be fitted on to a roadside pole. Claimed to be capable of capturing images from 50m away, Dtect can also count the number of people in a vehicle travelling at a maximum 80mph.
It works by projecting two low-intensity infrared (IR) beams of different wavelengths on to a vehicle windscreen. Using these wavelengths, Dtect takes two digital pictures of the vehicle, which are processed to produce one enhanced 'image' of the occupants' skin, discarding any non-facial features.
Protects anonymity
Software isolates the faces and counts them, and the image can have details such as a timestamp, location and any other information that the user may require imprinted on it.
Dtect preserves the anonymity of the driver by applying green dots to represent a face. 'We chose non-distracting IR wavelengths that the eyes cannot see, so drivers are not blinded at night,' said VOL director Dr John Tyrer. Although Dtect is only able to image through the front windscreen, he claimed it is able to count the number of people sitting in both the front and rear seats.
'You can have slightly tinted windscreens, but it would be breaking the law in many countries to black out the screen as you must have a certain transmission level in it and the front windows,' he said. Although the developers have not tested the device on vehicles with blackened dividers that would conceal rear-seat passengers, they have found that the IR waves will still penetrate mirrored windscreens, said Dr Leon Lobo, who worked on the technology.
'On some new cars now, particularly German and French ones, there are metallic windscreens that reflect heat,' he said. 'But the wavelength we have chosen will go through these.'
By targeting only human skin, the device rejects other materials such as hair, clothing or car upholstery. It is also able to distinguish between different parts of the body and the face. 'If, for example, somebody puts his or her left hand up to pretend it is a face, we reject it on the basis of the aspect ratio if a hand is held up with closed fingers; size if there is a clenched fist; or the number of protrusions created by gaps in between the fingers,' said Lobo. 'The closest thing is if two hands are put together — that has the right aspect ratio and size, but then the motorist can't drive.'
According to Tyrer, flat-rate congestion charging methods of reducing road traffic, such as those employed in London, is not a great enough incentive for people to use fewer cars.
'In 2002, the average speed was 6.9mph. It peaked just after the congestion charge began, to about 11mph, and now it is down to about 6.7mph, so it has not worked,' he said. 'If we cannot get people out of their cars on to public transport, we have to try and look at controlling the residual flow of vehicles.'
Tyrer suggests various methods of control, including charging single occupants of cars more, and opening up bus lanes to allow multiple-occupancy vehicles to reach their destinations quicker.
Dtect costs £96,000 a unit, and can be rented for trials at £3,000 a week.
英國鐵路公司如何推動凈零排放
Need a citation for that comment ´trains are at least 12 times more efficient than planes´. This greenhouse gas emission report rigged the figures by...