Optimization of Dual Frequency-Difference MIT Sensor Array Based on Sensitivity and Resolution Analysis

Magnetic induction tomography (MIT) is a promising technology for intracranial hematomas imaging. The existing problems in MIT are mainly low sensitivity and low resolution of the sensor array, which is necessary to be optimized. A number of optimizations exist to enhance the resolution of MIT, but so far, studies only considered the eigenvalue or singular value of the sensitivity, or the magnetic flux with noise-free data or low level of noise. In this paper, the optimization considers the intensity and uniformity of the sensitivity matrix with added noise. In order to enhance the resolution of dual frequency-difference MIT, the coil parameters optimization is studied by simulations from three aspects: the inner radius, the shape, and the number of turns. After the sensitivity analysis, the absolute dimensions of the optimal coil structure are determined. Based on the optimal structure, the brain hematoma is simulated with a human brain model and finite element method. Images are reconstructed from the modeled data with a 26-dB signal-to-noise ratio noise contaminated. The results indicated that the optimal sensor array can achieve the conductivity resolution of 0.25 S/m and the spatial resolution is 12.38% of the brain radius.