Tomography

Tomography is the method that underlies medical scanners, which are mostly large, fixed installations, e.g. for X-ray CT scanning and Magnetic Resonance Imaging (MRI). Fundamentally, the portability and adaptability of any tomography system depends on the nature of the measurement process that it exploits, and it turns out that many tomography systems are "agile" in these respects

Research into Agile Tomography by ERPE Academics applies the tomographic method to many different problems by extending the technology of the method using:

  • Novel low-noise electronic systems
  • Innovative image reconstruction techniques

ERPE research in Tomography follows three major threads:

  1. Chemical Species Tomography

    The measurement approach used here is spectroscopic absorption by particular target molecules, and the reconstructed images show the concentration distribution of the target species. The technology used is opto-electronic in nature, and our systems to date have exploited absorption spectroscopy in the near-infrared, where the communications industry has fostered the development of appropriate diode laser light sources, fibre-based optical components, and photodiodes.

    The unique IMAGER system provides images of the in-cylinder distribution of hydrocarbon fuel vapour in an operating automotive engine. The example below shows the in-cylinder fuel distribution from a sequence taken at 3,000 frames per second (fps) during a single compression stroke of a multi-cylinder engine running at 1500 rpm; a new image is taken every 3 degrees of crank angle rotation (°CA), and the sequence starts at 42°CA BTDC (before top dead centre).

    As well as showing the increasing average fuel concentration as the piston travels up the cylinder (i.e. colours becoming more red, rather than blue), the sequence indicates the rapid changes in fuel distribution that are unfolding in the cylinder. This information can be used in the development of new engines and fuels.

    Many other species are accessible, such as water, CO and CO2, covering a wide range of concentration. Currently, the method is being adapted for aerospace application in the £2.7 Million FLITES project (Fibre-Laser Imaging of gas Turbine Exhaust Species), with the Universities of Manchester, Southampton, Strathclyde and Stanford, in collaboration with Rolls-Royce, Shell, Covesion, Fianium and OptoSci.

  2. Electrical Impedance Tomography

    This method is applied to image the distribution of electrical conductivity within a subject, and the subject could be one of several types of engineering process (e.g. multi-phase flow), or a medical patient undergoing intensive care (e.g. for respiratory problems). Typically, current is injected through the subject from different angles in a sequence, and the resulting voltage distribution for each current pattern is measured on the boundary of the subject.

    For appropriate measurement speed and sensitivity, state-of-the-art digital low-noise electronics is used, as illustrated in the FPGA-based medical EIT system.

  3. Image Reconstruction Algorithms

    There is considerable complexity and challenge in the process of inverting tomographic data to create images of the distributions which gave rise to the set of measurements, and this is an example of a mathematical inverse problem. In particular, we are typically dealing with small numbers of measurements, a few dozen, compared with numbers of about a million in medical systems.

    This has led to a number of innovative image reconstruction algorithms, as illustrated by the CST examples below: the diagram on the far left shows the dimensions of a phantom (i.e. already known) distribution of propane gas produced in a bench-top rig in the laboratory; using the experimental measurements on this phantom, the reconstructed distributions shown 2nd and 3rd left illustrate the results of applying different filters during the iterative Landweber reconstruction process.

Research Partners

School of Engineering, University of Edinburgh:

  • Hugh McCann (UoE)
  • Nick Polydorides (UoE)
  • Jiabin Jia (UoE)

Industrial Partners

 

Research Institutes: 

Research Themes: