Enhancement of microwave tomography using Kalman filter theory

Abstract

A strategy of microwave tomography enhancement using Kalman filter theory is proposed. A known object was placed in the imaging domain, and its inversion result was compared with its actual value to extract prior information. The differences between actual value and inversion result were due to the low pass filter performance determined by the microwave tomography system. Singular value spectrum of integral operator, which bases on the Green's function, was analyzed to show the low pass filter fact. Then we used a prior information obtain from known object to build a Kalman Gain Matrix, this matrix is used to predict and update the inversion result of target, since the inversion result was processed by the same low pass filter as the known object one. The synthetic study shows that the Kalman filter theory can be used in microwave tomography to deal with its inherent low pass filter performance, and leads to more accurate results.