An Improved Tikhonov Method for Magnetic Induction Tomography

Magnetic induction tomography (MIT) is a non-contact and more advanced technique than CT imaging. The main reason why it is still in the laboratory research stage are that the edges of reconstructed image are blurred and cannot satisfy quality requirements. Based on this, a random boundary optimization method is proposed for MIT. The method is to optimize edge information. Firstly, we chose the Tikhonov optimization algorithm with poor boundary information, the image reconstruction result of Tikhonov is seen as the initial state. Then, through the K-means clustering theory, a binary image is produced; By using the expansion and corrosion method of mathematical morphology, the boundary information is extracted; Particle swarm optimization is used to optimize the boundary information. Finally, experiment result shows that the image error and correlation coefficient are much better than Tikhonov algorithm. So it is a kind of MIT image reconstruction method with high accuracy.