Fast Baggage Scanning: Mathematics Crucial to Airport Baggage Screening
By Prof Bill Lionheart (University of Manchester)
In 2005, a Surrey University spin out company CXR approached us at the University of Manchester for help developing a fast x-ray tomography system. Coming from a perspective of highly non-linear and ill-posed imaging problems my first (and naive) reaction was x-ray CT (Computed Tomography) was quite simple . What drew me in to the problem was the novelty of the geometry. A conventional hospital x-ray CT system has an x-ray source and array of detectors mounted on a mechanically rotated gantry, and this limits its rotation and hence scanning speed. The Real Time Tomography (RTT) system developed by Ed Morton and his team had a fixed array of x-ray sources which were electronically switched. The down-side is that the detector array has to be offset from the ring of sources rather than opposite, and there was no theory nor reconstruction algorithm for for this case.
The original idea of the RTT was to monitor oil, water and gas in pipes in the petroleum industry but CXR was bought by the major security screening company Rapiscan Systems, in the wake of an increased interest in improved baggage screening following 9/11, with the intention of developing a fast 3D baggage scanner. In a joint EPSRC and Rapiscan funded project worth £750k over three years we developed novel reconstruction algorithms, optimal measurement schemes and scatter correction algorithms for the RTT, which were patented and published. By the end of the project there was a production prototype baggage scanner. After regulatory approval the RTT 110 went in to production at its UK factory and is now deployed in airports around the world scanning hold luggage on a conveyor belt in real time.
At Manchester the work widened our interest in x-ray CT algorithms especially for materials and advanced manufacturing in collaboration with the Henry Mosley X-ray Imaging Facility. We now have our own RTT 110 which is being tested and adapted to do fundamental science such as looking in to the mysteries of granular materials and two phase fluid flow, as well as rapid non-destructive testing for manufacturing.
Interdisciplinary projects are always more challenging, mainly due to communications barriers. This is true even more so in collaborations between industry and academia ,but if you meet these challenges, the rewards are worth it.
Prof Bill Lionheart is Professor of Applied Mathematics, University of Manchester. He specialises in inverse problems and imaging across a wide range of applications including security screening, medical imaging, non-destructive testing, and land mine detection. EPSRC Grant Information: EP/E010997/1 Title: X-ray CT reconstruction algorithms for airport security and process tomography. This was a special maths with industry EPSRC maths programme call. 500k from EPSRC and 250 k from Rapiscan