62
thus increasing image acquisition time. Traditional pulse THz systems relies on the
slow mechanical movements of the delay line to scan a time-domain waveform. Several
faster alternatives exist such as asynchronous optical sampling [31,32] and single-shot
chirp-pulsed detection [36-38]. If interested only in imaging phase information at one
frequency, table-top continuous-wave systems based on diodes and frequency multi-
pliers from Virginia Diode, Inc., which operates at 0.1THz to above ITHz, can be
used to build CS imaging systems. Each of these systems have their advantages and
drawbacks, which should be carefully considered before the implementation of a CS
imaging system.
Last but not least, the ongoing research in the area of compressive sensing will
continue to inspire new imaging techniques and applications [95]. For example, CS
has only been used to reconstruct mostly 2-D THz images in this thesis. Further
research in a third dimension, such as CS reconstruction of THz spectroscopic im-
ages or imaging∕detection of cracks using lactose powder near its THz absorption
frequency [96], will further enhance the imaging capability of future CS THz imaging
systems. A great example of the application of CS to 3-D THz data is demonstrated
in the context of pulse-echo mode THz reflectance tomography [97]. Development of
reconstruction algorithms for complex THz data or THz phase images will also be
important for CS applications in 3-D [72,73].
To conclude, this thesis is a first step to bridge the gap between the theory of
compressive sensing and terahertz imaging. This piece of research, hopefully, will in-
spire future research in applying advanced signal processing theories to build practical
imaging systems!