Tunnel vision

London Underground is to trial a digital imaging system that monitors the movement of train tunnel walls.

A full prototype of the technology was first used last spring by engineers on the Channel tunnel Rail Link (CTRL) to monitor the effect the new tunnel's construction might be having on the stability of the North London rail line above it.

According to Keith Bowers, who worked at the time for consultant Rail Link Engineering but is now employed by London Underground, his new employer is set to use the technology towards the end of the year.

Conventional surveying techniques such as robotic total stations or manual precise levelling are often used in projects such as the CTRL to measure the displacement of walls and embankments during tunnel boring.

While these can provide an extremely high level of precision, there are occasions when it is not always practical to use them. This can be due to frequently passing trains, an electrified track or where it is too dangerous to allow surveyors easy access to the site.

The new system - named HIMY (Has It Moved Yet?) based on a surveyors' in-joke - uses a technique known as Particle Image Velocimetry (PIV) to calculate wall movement. The technology uses a standard digital camera and PC in a reinforced box positioned opposite the area under consideration. Every 30 minutes a picture is taken and beamed back in real-time via a GPRS modem to a remote server.

A digital image is a matrix which contains the brightness recorded at each pixel on the Charge Coupled Device of the camera. By comparing the brightness of this matrix taken at two different times, PIV's software 'learns' the unique pattern of pixel intensity in the area, and then searches for a section of wall which is of identical size and with a matching set of intensity matrices for every pixel.

By doing this, the movement of that section of wall - however slight - can be monitored as the movement in pixels can then be translated into more useful engineering units.

The software automatically calculates the displacement of the wall section and adds each image to a database which shows how much the section has moved over time. Bowers said that the fact that the system uses images as opposed to just pure data is useful for a number of reasons.

'First, it combines images with data that makes it easier for an engineer to look at and quickly understand without having to trawl through lots of data,' he said. 'Second, because each image is on the database, we can look retrospectively at the images, so if a defect is spotted we can go back and measure its development over time, even if we hadn't known it was there at the time.'

According to Bowers, HIMY's performance is comparable in accuracy to conventional surveying systems and that this accuracy depends on the number of pixels that the camera can capture. 'We can use a different lens and have a higher resolution of a smaller field of view and vice versa,' he said.

The technology is still at an early stage, and Bowers noted that as the pixel-capturing capabilities of digital cameras improve, the system's accuracy could also be greatly enhanced in the future.