Abstract:

In addition to already established imaging techniques like magnet resonance imaging (MRI) or computed tomography (CT), microwave imaging systems have been developed recently and their usage for biomedical applications e.g. stroke detection is researched. Microwave imaging is a promising technique which is expected to lead to a portable and cost effective detection system.For image reconstruction in this area there exists a wide range of approaches and algorithms from radar approaches like beamforming algorithms to tomographic approaches which are based on optimisation problems.An important part of this development is the adaption and optimisation of algorithms used for microwave imaging.In this thesis, a software architecture of an object-oriented framework, which can be used for development of different algorithms, was developed.In addition to the algorithm development, the main purpose of the framework is to provide the opportunity to compare different algorithms of microwave imaging techniques based on equal input data or image size. For an easy extensibility of the framework a modular concept was developed which also respects the requirements of parallel computing of algorithms on a computation cluster.To proof the useability of the software architecture for microwave imaging in the context of stroke detection, different signal processing techniques are integrated into the framework.The basic image reconstruction algorithm used is a beamforming algorithm based on the delay-and-sum principle.This algorithm uses ultra-wideband signals calculated through a finite-difference time domain simulation of a three dimensional head phantom model. Both the reconstruction results and the perfomance of the software architecture was analysed with respect to the scalability of the developed concept.