“Hard-field” Temperature Imaging by Guided Path Tomography

S. G. Castillo and K. B. Ozanyan

School of Electrical and Electronic Engineering, The University of Manchester, Manchester, M60 1QD, UK, Email: k.ozanyan@manchester.ac.uk

ABSTRACT

Temperature imaging with Guided-Path Tomography (GPT) is based on the measurement of electrical signals from a small number of non-interacting transducers (wires or discrete sensors), yielding line integrals of the temperature-dependent resistivity. GPT was introduced at the 3rd WCIPT (Ozanyan, 2005) and the feasibility of imaging temperature on non-planar surfaces, using the Additive Algebraic Reconstruction Technique, has been demonstrated (Al-Shaikh, 2005). Here, we report on the possibility of obtaining strictly Radon-transformed data from electrical measurements and consequently reconstruct the temperature field by inverse Radon transform, using Filtered Backprojection (FBP) or Direct Fourier Inversion.

In this work, the GPT temperature sensor (Al-Shaikh, 2005) is configured as a set of 32 parallel nickel wires, analogous to 32 pencil beams in hard-field tomography, yielding the imaged subject’s projection at a certain angle. The Radon transform of the temperature field is sampled by a 180o mechanical rotation of the sensor head and recording 60 sets of projection values. As the room temperature resistance of the transducers is in the order of 1Ω to ensure maximum temperature sensitivity, 32 identical front-end circuits are used to connect each transducer to a dedicated FPGA-based digital system (Garcia Castillo, 2007), allowing 32 parallel (“projection-at-once”) data streams. The individual resistivity measurements are taken with an internally generated 180 kHz ac signal, modulated by the resistance changes induced by the temperature variations. 32 channel lock-in detection (Garcia Castillo, 2005) is simultaneously performed at each transducer and the processed data are transferred through a dedicated interface to a PC. There the image reconstruction is performed and displayed using LabVIEW software with embedded MATLAB FBP routines.

Keywords Guided-Path Tomography, Temperature, Digital Tomograph, ΣΔ ADC, FPGA, DSP.