Microwave imaging in dispersive media using time reversal techniques in high performance computing environment

• Main Author: Abduljabbar, Ammar Muwafaq
• Other Authors: Costen, Fumie
• Published: University of Manchester 2018
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.740369

The Time Reversal (TR) techniques achieve spatio-temporal refocusing either by physical or synthetic retransmission of signals acquired by a set of transceivers in a time-reversed fashion which can be used in various applications, including microwave imaging of hidden targets. This is due to the invariance of wave equations in lossless space. However, the existence of dispersion and loss in the propagation medium breaks this invariance and the resultant TR focusing exhibits frequency and time-dependent degradation. Compensation methods can tackle this degradation to improve focusing resolution under such conditions. In this thesis, we propose an algorithm that utilises inverse filters with threshold approach and different type of windows to compensate for this additional attenuation. The proposed threshold approach reduces the amplification of the unwanted noise in the received signals at the application of the inverse filters. Furthermore, optimum settings for window type and length in the Short Time Fourier Transform (STFT) method are obtained through a scanning operation in the propagation medium. While utilizing a large number of windows with short spatial lengths provides improved TR focusing performance, it also increases the overall cost and complexity of the imaging system. The threshold method introduced here achieves improved TR focusing performance without increasing the cost by utilising a lower number of inverse filters. We also identify the limitation of the STFT compensation method in some human tissues. The limitation is overcame by our proposed compensation method with the Continuous Wavelet Transform (CWT) technique. The CWT method uses a size-adjustable window while the STFT method uses a fixed window. This thesis investigates the performance of conventional TR, TR MUltiple SIgnal Classification (TR-MUSIC) and total focusing techniques to detect a tumour in the lung after application of the CWT compensation method to the observed signals.