Filtered forward-backward time-stepping method applied to reconstruction of dielectric cylinders

In this paper, an optimization approach for reconstructing the electrical parameters of cylindrical objects from total field data in the time domain is examined. Noise in measured field data degrades the quality of reconstructed image based on the forward-backward time-stepping (FBTS) algorithm previously proposed. In order to improve the image quality, we propose a new reconstruction algorithm. The measured time domain data are low-pass filtered and the FBTS algorithm is used for the filtered data. Numerical simulations are carried out to show the effectiveness of this algorithm for a high-contrast object.