Ultrasound tomography: measurement setup and reconstruction methods
Daniel Eder
Ultrasound imaging based on the tomographic principle is gained interest for medical and technical applications in recent years. The goal of this thesis is to build an ultrasonic scanner from simple components to be able to acquire data on various test objects for an investigation on the use of different tomographic reconstruction methods. With the inverse fan-beam transform known from computed tomography and an iterative reconstruction approach based on discretized Bézier curves, two different reconstruction methods are applied to the data. Both aim at the reconstruction of the two-dimensional speed of sound distribution in the measurement volume.
For the generation and detection of ultrasonic signals, a total of 16 piezoelectric ultrasonic transducers are used in the measurement setup. As part of this thesis, the sound field of these transducers is characterized very precisely and the impact of 3D-printed acoustic lenses on their sound field is investigated. The insights of this investigation are used to determine a suitable geometry for the arrangement of the transducers in the measuring head. A cylindrical diverging lens is applied to the emitting transducer of the scanner to achieve a wider radiation pattern.
In the course of the work, the scanner is tested using two different objects. To use one inorganic and one organic test object, a homogeneous polyamide cylinder and a boiled and peeled hen's egg are used. It is shown that the chosen reconstruction algorithms reconstruct the position and shape of both objects well and thus provide plausible results. However, reconstruction results with simulated measurement data show that the setup still has significant potential for improvement. Increasing the number of receiving transducers, for example, would enable a much more accurate reconstruction of the speed of sound distribution.
Keywords: ultrasound tomography, fan-beam-transform, iterative reconstruction, ultrasound imaging, measurement technology, ultrasound
13th January 2022