Many materials such as drugs and explosives have characteristic spectral signatures in the terahertz (THz) band. These unique signatures imply great promise for spectral detection and classification using THz radiation. While such spectral features are most easily observed in transmission, real-life imaging systems will need to identify materials of interest from reflection measurements, often in non-ideal geometries. One important, yet commonly overlooked source of signal corruption is the etalon effect - interference phenomena caused by multiple reflections from dielectric layers of packaging and clothing likely to be concealing materials of interest in real-life scenarios. This thesis focuses on the development and implementation of a model-based material parameter estimation technique, primarily for use in reflection spectroscopy, that takes the influence of the etalon effect into account. The technique is adapted from techniques developed for transmission spectroscopy of thin samples and is demonstrated using measured data taken at the Northwest Electromagnetic Research Laboratory (NEAR-Lab) at Portland State University. Further tests are conducted, demonstrating the technique's robustness against measurement noise and common sources of error.
| Identifer | oai:union.ndltd.org:pdx.edu/oai:pdxscholar.library.pdx.edu:open_access_etds-1240 |
| Date | 01 January 2010 |
| Creators | Kniffin, Gabriel Paul |
| Publisher | PDXScholar |
| Source Sets | Portland State University |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Dissertations and Theses |
Page generated in 0.0021 seconds