Research On Multi-frequency Data Acquisition Technique For Electrical Impedance Multi-frequency Of Parameter Tomography

Multi-frequency electrical impedance tomography (MFEIT) and electrical impedance parameter tomography (EIPT) are newly developed technique based on electrical impedance tomography (HIT). They can acquire the information of the distribution or the changes of bio-impedance, and reconstruct the impedance images at several different frequencies. So, the reconstructed images not only have the characteristics of functional imaging as conventional HIT has, it also has the features that the stimulating frequency can be selected to improve the impedance contrast of the target tissue from others. More over, it can extract bio-impedance characteristic parameters from those images to build parameter images, which may be valuable to identify tissues and evaluate its states. So, MFEIT and EIPT are a new trend in EIT research, and have good prospects in clinical application.In order to overcome the shortcomings in sweeping frequency system and mixing frequency system, which are commonly used presently. We developed a new 32-electrode data acquisition system (DAS) for multi-frequency EIT based on frequency mixing and frequency sweeping technique, we also set up a software platform for the system. The DAS has four working modes, and can simultaneously collect impedance information at four different frequencies. The frequencies can be swept from 1.6kHz to 380kHz under the control of our software.In order to improve the data acquisition speed and precision, the following methods have been studied:1.Two independent measurement channels, which can acquire theimpedance signal synchronously, are used to improve the acquisition speed.2.In order to use a general data acquisition card to sample high frequency signal efficiently, an improved sampling method is developed.3.A new demodulation method named by us as "orthogonalsequential digital demodulation" has been developed. It has theadvantages of high calculation speed and high precision. For asignal with n components of frequency, using this method andsampling N points per period, the total needed operations can bereduced to 2nN multiplication and 2n(N-l) addition operations,the signal to noise ratio can be improved by (TV/2)0-5 times.In order to get the EIPT images, an algorithm based on the least meansquare method is developed to extract the bio-impedance characteristicparameters from the multi-frequency static EIT images or quasi-static HITimages.From the measured results of hardware system, we can see that at operational frequency range, for small signal about 1 mV(RMS) the SNR is above 50dB. For large signal about lOOmV(RMS) the SNR is large than 80dB. Under the control of Windows message queue, it needs two seconds to acquire impedance signal to reconstruct a frame of impedance image.The dynamic images and quasi-static images based on physical phantom also show the corresponding impedance changes in phantom or resulted from the changed frequencies.Finally, we analyzed the performance of the system and discussed the possible methods for improvement in the further research.