DYNAMIC MICROWAVE IMAGING

This dissertation extends the techniques of microwave diversity imaging, which has been established and proven to be capable of yielding high resolution microwave imagery of various test objects under stationary conditions, to dynamic imaging environments. The objective is to undertake a careful and thorough analysis of all the factors involved that would enable realizing in dynamic imaging the same unprecedented image resolution and quality achieved in earlier imaging of stationary objects. The theory, principles and techniques for microwave stationary diversity imaging are reviewed, and some of the concepts clarified, to set the stage for the dynamic imaging analysis. In particular, angular diversity is discussed with the aid of the Ewald sphere representation, showing several modes of scanning arrangements, together with frequency diversity, to access different parts of the Fourier space and enable projective, tomographic, and three-dimensional imaging. The concept of accessing a part of the Fourier space is utilized to analyze image formation of various dynamic imaging scenarios. Different illumination waveforms, and how to wrest Fourier information out of the object response to these waveforms, are addressed. The procedures entailed in correcting the measured data and obtaining the Fourier space samples are discussed. In particular, the basis of the range alignment and reference range inference essential to image retrieval is examined and algorithms to achieve them are proposed and verified. The tolerances under which these algorithms can be applied to alignment of the range profiles and to determine the reference range are presented. The effects of angular errors and frequency errors on the images are discussed in terms of polar format sampling errors. Also given are the criteria such that these errors would not inflict too serious image degradation. Finally, the problems of the presence of gross Doppler in a dynamic imaging situation are addressed. It is shown that there are circumstances that this gross Doppler can be ignored, e.g. when the illumination consists of short bursts of pulses. However, if the illumination time is long, a measurement scheme called orchestrated Doppler compensation measurement is proposed to nullify the Doppler effect. (Abstract shortened with permission of author.