Extraction of anthropological data with ultrasound

Abstract

Human body scanners used to extract anthropological data have a significant drawback, the subject is required to undress or wear tight fitting clothing. This thesis demonstrates an ultrasonic based alternative to the current optical systems, that can potentially operate on a fully clothed subject. To validate the concept several experiments were performed to determine the acoustic properties of multiple garments. The results indicated that such an approach was possible. Beamforming is introduced as a method by which the ultrasonic scanning area can be increased, the concept is thoroughly studied and a clear theoretical analysis is performed. Additionally, Matlab has been used to demonstrate graphically, the results of such analysis, providing an invaluable tool during the simulation, experimental and results stages of the thesis. To evaluate beamfoming as a composite part of ultrasonic body imaging, a hardware solution was necessary. During the concept phase, both FPGA and digital signal processors were evaluated to determine their suitability for the role. An FPGA approach was finally chosen, which allows highly parallel operation, essential to the high acquisition speeds required by some beamforming methodologies. In addition, analogue circuitry was also designed to provide an interface with the ultrasonic transducers, which, included variable gain amplifiers, charge amplifiers and signal conditioning. Finally, a digital acquisition card was used to transfer data between the FPGA and a desktop computer, on which, the sampled data was processed and displayed in a coherent graphical manner. The beamforming results clearly demonstrate that imaging multiple layers in air, with ultrasound, is a viable technique for anthroplogical data collection. Furthermore, a wavelet based method of improving the axial resolution is also proposed and demonstrated.