Signal processing for noise equalisation within EIT images

The Sheffield Mk3 system is a multi-frequency EIT (electrical impedance tomography) system with up to 32 parallel receive channels and 32 drivers. It uses up to 64 electrodes with a constant current drive and wide-band, high input impedance voltage measurement on alternate adjacent pairs of electrodes in the frequency range 9.6 kHz to 1.2 MHz. Many of the changes of interest appear towards the centre of the images and the changes may be less than 1%. Unfortunately, the accuracy of the measurements in these regions is poor. It is desirable to improve the system performance by increasing the SNR in the central area of the images whilst maintaining an adequate SNR in the peripheral part of the image. This paper presents a noise equalisation method, that is based on the use of frequency filtering but with a narrower bandwidth applied to the small signals than to the large ones. This method of noise equalisation can be applied to both serial and parallel data collection-based EIT systems